Ultrastructural localization of prolactin and chromogranin B messenger ribonucleic acids with biotinylated oligonucleotide probes in cultured pituitary cells.

A technique for the subcellular localization of prolactin and chromogranin B messenger RNAs (mRNA) in pituitary adenomas by in situ hybridization with biotinylated oligonucleotide probes is described. Ultrastructural examination revealed clusters and individual gold particles in the cytoplasm with this pre-embedding in situ hybridization method. Prolactinomas expressed both prolactin and chromogranin B mRNA, whereas the null cell adenoma expressed only chromogranin B mRNA. Sections of positively labeled cells contained up to 30 gold particles/cell. Treatment of cells with RNAse before hybridization reduced the number of gold particles to less than 1/cell. These results indicate that biotinylated oligonucleotide probes can be used to localize different mRNAs in cultured pituitary cells at the ultrastructural level to study the processing of these molecules within specific cells and for more precise correlation of molecular function with ultrastructural morphology.

Characterisation of tissue-specific trans-acting factor binding to a proximal element in the rat growth hormone gene promoter.

Using an exonuclease III protection assay, tissue-specific binding of rat pituitary tumour cell (GH3 cell) nuclear factors to a proximal region (-68 to -138) of the rat growth hormone gene promoter has been detected. The binding is particularly strong between the borders -68 to -102. The binding is eliminated in the presence of excess unlabelled rat growth hormone gene promoter sequences but also by proximal (-423 to +38) or distal (-1960 to -1260) rat prolactin gene promoter sequences and simian virus 40 enhancer/promoter sequences. Extracts of rat pituitaries showed identical binding characteristics. Methylation interference analysis indicated that the contact points between the pituitary-specific factor and the proximal rat growth hormone gene promoter-binding element (-65 to -95) are over a conserved sequence which occurs twice in the rat growth hormone gene promoter and at least eight times in the rat prolactin gene 5'-flanking sequences. This sequence has previously been proposed to constitute the binding site for the somatotroph/lactotroph tissue-specific transcription factor. Gel-retardation and exonuclease III competition analysis showed that three of the rat prolactin gene promoter elements (-46 to -71, -156 to -180 and -174 to -204) share the ability to bind the pituitary-specific factor. The binding to the most proximal rat prolactin gene promoter element (-46 to -71) was clearly more avid than to the rat growth hormone gene promoter (-65 to -95) proximal element. However, both these elements displayed the formation of two gel-retarded complexes while the more distal rat prolactin gene binding elements (-156 to -180 and -174 to -204) formed only the smaller of the two complexes. Finally, we demonstrated by co-transfection competition analysis that plasmids containing the most proximal rat prolactin gene promoter binding element completely inhibited transcription from the rat growth hormone gene promoter while rat growth hormone gene promoter sequences only partially inhibited transcription from the rat prolactin gene promoter. This suggests that the higher affinity for factor binding displayed by the proximal rat prolactin gene promoter binding site in vitro is reflected in factor binding activity in vivo.

Estradiol stimulation and inhibition of cell growth in new estrogen-sensitive cell lines and tumors established from the MtTF4 tumor.

Cell lines were established from the MtTF4 tumor, growth of which is inhibited by estradiol, in order to determine whether the effect observed in vivo was due to a direct action on tumor cells. Two different cell lines were obtained according to the medium in which tumor cells were dispersed and cultured. The F4P cells were obtained when the culture medium contained charcoal-treated fetal calf serum. The growth rate of these cells was slowed down by 17 beta-estradiol in animals and also in culture during the early passages. Thereafter, they became insensitive to 17 beta-estradiol in culture but remained negatively controlled in vivo. These cells, whatever their sensitivity to 17 beta-estradiol, secrete prolactin and carry functional D2 dopamine binding sites. The F4Z cells were established in medium containing fetal calf serum not treated with charcoal. The growth rate of these cells was stimulated by 17 beta-estradiol in animals but was 17 beta-estradiol insensitive in culture up to subculture 26. At this time, the growth rate of the subline also became stimulated by 17 beta-estradiol in culture, and this phenotype was still found at passage 108 (50% effective dose, 5 to 10 pmol; maximum stimulation, 180 to 300% of control). These cells neither secrete measurable amounts of prolactin nor have dopamine binding sites. Thus, according to the medium in which cells were dispersed and cultured, two different cell strains were derived from a tumor in which growth is inhibited by 17 beta-estradiol. The point of interest is that the growth rate of one strain was inhibited by 17 beta-estradiol, while the other was stimulated. Convergent data suggest that MtTF4 tumor was heterogeneous and that selection had occurred during the dispersion or the culture of cells. Since the growth of one of these cell lines was slowed down transiently in culture we conclude that the inhibition of tumor growth could be due to a direct action of 17 beta-estradiol on tumor cells. However, the dissociation between the response to 17 beta-estradiol in culture and in the animal observed at some time of cell evolution suggests that environment affects the sensitivity of cells to 17 beta-estradiol.

Expression of a cytomegalovirus-human growth hormone-releasing hormone precursor fusion gene in transfected GH3 cells.

Pituitary GH3 cells were transfected with a human growth hormone-releasing hormone (hGRH) precursor minigene fused to the promoter region of either a cytomegalic immediate early gene (pCMV) or the mouse metallothionein-1 gene (mMT) to examine the molecular heterogeneity of the translation products. Expression of the hGRH message occurred following transfection of the cells with each fusion gene. Extracts of pCMV-hGRH-transfected GH3 cells as well as the culture medium contained detectable levels of immunoreactive (ir)-hGRH peptides. Analysis of molecular heterogeneity by reverse-phase high performance liquid chromatography and radioimmunoassay indicated that both mature forms of hGRH (hGRH(1-44)-NH2 and hGRH(1-40)-OH) were synthesized in the cells, although hGRH(1-44)-NH2 was the primary form secreted into the medium. A high molecular weight form of ir-hGRH, believed to represent the hGRH precursor (or a partially processed form of the precursor) was detected in cells and, in smaller amounts, in the medium. Several ir-hGRH peptides, presumed cleavage products of the mature forms of hGRH, were also found. The efficiency of processing of the hGRH precursor and metabolism of the mature hormonal forms in transfected cells grown in the presence of four different peptidase inhibitors varied with the inhibitor present. Transfected GH3 cells, therefore, possess all of the necessary enzymes for and are capable of processing the hGRH precursor to mature GRH and provide a model to study hGRH biosynthesis.

Characterization of a tissue kallikrein in human prolactin-secreting adenomas.

Immunoreactive tissue kallikrein was co-localized with prolactin in all the eleven prolactin-secreting adenomas of the human anterior pituitary gland examined in this study. The intracellular distribution of immunoreactivity in the prolactin-secreting cells suggests that tissue kallikrein is located within the Golgi complex of these cells. Both the intracellular hormone-processing action and the kininogenase activity of tissue kallikrein may be of functional importance in human prolactinomas.

[An autopsy case of primary intracranial squamous cell carcinoma].

An autopsied case of primary intracranial squamous cell carcinoma (PISCC) is reported, and 25 previously reported cases of PISCC, followed by the Garcia's criteria, are reviewed. A 72-year-old female was admitted to our service with chief complaints of headache and nausea on March 30, 1988. She had no neurological deficits on admission. However, CT examination revealed a round mass lesion in the left hypothalamus with dislocation of the brain stem. The cerebrospinal fluid (CSF) examination showed squamous cell carcinoma cytologically, and slightly higher levels of beta-HCG (13.0 ng/ml) and CEA (14.2 ng/ml). Because of progressive worsening in the level of her consciousness, total removal of a suprasellar tumor was performed on April 19, 1988. Gross appearance of the tumor was yellowish, soft and encapsulated. Histologically, it was squamous cell carcinoma. She did well for several days after the operation, then deteriorated. Finally she expired because of dissemination of the carcinoma on May 14, 1988. Postmortem examination revealed a large mass of squamous cell carcinoma in her right cerebellopontine angle. Except for that in the brain, no cancer was found in her body. Immunohistological study of the tumor specimen demonstrated positive for HCG in some of the large-sized neoplastic cells. Twenty-six cases of PISCC have been reported previously, so far. However, 21 cases out of the 26 PISCC were thought to have originated from intracranial epidermoid, one from the dermoid and the other one from craniopharyngioma. In the other three cases of PISCC, including the present case, the origin of the tumor was not able to be identified.(ABSTRACT TRUNCATED AT 250 WORDS)

Intermediate filament expression in pituitary adenomas.

Seventy-five formalin-fixed and 18 alcohol-fixed pituitary adenomas were studied immunohistochemically using antibodies to keratin, vimentin, neurofilaments (NFs), glial fibrillary acidic protein, desmin, actin, S-100 protein and a variety of pituitary hormones. The pituitary adenoma cells were positive for keratin, vimentin and NFs (68 kDa and 160 kDa) and in a few instances there was co-expression of these three types of intermediate filaments (IMFs). The pattern of keratin-specific staining showed diffuse cytoplasmic or patchy paranuclear reactivity and of NF- or vimentin-specific staining showed fibrillar or patchy paranuclear reactivity. The patchy staining seemed to decorate the fibrous body. There was no correlation between the distribution of IMFs and pituitary hormones in pituitary adenomas except that melanocyte-stimulating-hormone-positive reactivity was limited to the NF-positive adenomas. The pattern of IMF staining did not depend on hormone production in adenomas.

Pituitary pathology in acquired immunodeficiency syndrome.

Pituitary morphology was studied in 49 autopsied patients with acquired immunodeficiency syndrome. Direct infectious involvement was noted in six adenohypophyses (12%), including five cases by cytomegalovirus and one by Pneumocystis carinii. Two cases with neurohypophysial lesions presumably caused by cytomegalovirus and one questionable case of Toxoplasma gondii were also observed. In all instances these changes were associated with generalized and/or cerebral infection by these same agents. Neither Kaposi's sarcoma nor malignant lymphoma was encountered in the pituitary glands. Acute necrotic foci, presumably due to infarction, were noted in four cases. Four pituitary microadenomas (8%) and four hyperplastic nodules were identified. The incidence of such noninfectious lesions, as well as the prevalence and distribution of the various immunoreactive adenohypophysial cell types, were similar to those seen in the pituitary glands of age-matched male control patients.

DNA evaluation in growth hormone producing pituitary adenomas: flow cytometry versus single cell analysis.

DNA patterns were analysed in 26 GH-producing pituitary adenomas by flow cytometry as well as by microspectrophotometry. Twelve tumours (46%) were diploid according to both methods, whereas 5 tumours (19%) showed aneuploid DNA patterns. Nine tumours were classified differently by the two methods: flow cytometry resulted in diploidy in 2 and aneuploidy in 7 patients, whereas microspectrophotometry showed diploidy in 5 tumours, tetraploidy in 3 and aneuploidy in 1. Methodological limitations may explain the discrepancy in the results obtained by the two methods. However, both the flow cytometry and the microspectrophotometry method show the presence of aneuploid DNA patterns in GH-producing pituitary adenomas despite their benign growth characteristics and the clinically benign course of the disease. This comparative study with two methods measuring DNA content, shows that depending on the criteria used for diploidy-aneuploidy, the frequency of aneuploidy will vary. In this material of 26 GH-producing adenomas, 46% were aneuploid according to flow cytometry and 23% according to microspectrophotometric. However, no correlation to tumour size or GH levels was found with either method when patients with aneuploid and diploid tumours were compared. Therefore, no clinical significance can so far be drawn from these results.

Altered balance between thyrotropin-releasing hormone and dopamine in prolactinomas and other pituitary tumors compared to normal pituitaries.

We measured TRH and dopamine (DA) concentrations in prolactinomas and other pituitary tumors in order to further understand the roles of these two factors in the hormone hypersecretion and growth of these tumors. The mean TRH concentration (by RIA) in 16 prolactinomas was 247 +/- 92 (+/- SE) fmol/mg cell protein (range, 10-1297), near that found in normal pituitary tissue. The prolactinoma TRH content did not correlate with the patient's tumor size or plasma PRL level. By contrast, DA assayed by high pressure liquid chromatography was present in normal pituitary tissue (7.3 +/- 3.5 pmol/mg cell protein), but was very low or undetectable in the prolactinomas (23 fmol/mg cell protein or less). 3,4-Dihydroxyphenylacetic acid, also assayed by high pressure liquid chromatography, was undetectable in both normal pituitary tissue and prolactinomas. This imbalance between TRH and DA content also was found in GH-secreting and nonsecreting adenomas. The TRH content in 18 GH-secreting tumors (24 +/- 6 fmol/mg) was considerably lower than that in the prolactinomas (P less than 0.001). In 8 nonsecreting adenomas, the mean TRH concentration was 109 +/- 28 fmol/mg, about half of that in the prolactinomas. In those 2 types of adenomas, DA also was nearly undetectable (less than or equal to 73 fmol/mg cell protein). We conclude that the imbalance between TRH and DA contents in prolactinomas compared to those in normal pituitary tissue might participate in the mechanisms leading to hypersecretion of PRL and the growth of all types of pituitary adenomas.

Evidence of pre-prosomatostatin mRNA in human normal and tumoral anterior pituitary gland.

Expression of the SRIH gene was investigated in six human normal anterior pituitaries, six GH-, three PRL-, three mixed GH/PRL-secreting and four nonsecreting adenomas. Total cellular RNA and poly(A+) mRNAs were analyzed by dot and Northern blot hybridization to a 3'-end labeled oligonucleotide probe specific for the human pre-proSRIH mRNA. A weak but detectable pre-proSRIH hybridization signal was present in human normal anterior pituitaries and in the four groups of adenomas. The size of this pre-proSRIH mRNA was indistinguishable from that found in our hypothalamic samples and close to that described in the literature. The wide variation of the signal intensity from one case to the other in each group of the different types of normal and tumoral antehypophyseal samples prevented establishment of any correlation between the level of pre-proSRIH mRNA and the nature of the pituitary tissue. The presence of SRIH mRNA in human normal and tumoral anterior pituitary tissues provides a sound basis to substantiate the hypothesis of a SRIH biosynthesis in the human anterior pituitary gland.

Evidence for multiple protein constituents of the somatostatin receptor in pituitary tumor cells: affinity cross-linking and molecular characterization.

Previous studies have shown that somatostatin receptors on AtT-20 and GH3 pituitary tumor cells show relative preference for binding somatostatin-28 (S-28) and somatostatin-14 (S-14), respectively. Here we have attempted to determine whether this selectivity can be explained by molecular heterogeneity of the receptor. Cells were incubated with [125I-Tyr11]S-14, [125I-Leu8-D-Trp22,Tyr25]S-28, and [125I-Tyr3]SMS, and the bound ligand was chemically cross-linked with bis-[2-succinimido-oxycarbonyloxy)ethyl]sulfone, disuccinimidyl suberate, or dithiobis (succinimidyl propionate). The solubilized cross-linked material was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by autoradiography. [125I-Tyr11]S-14 labeled three specific receptor proteins of 57K, 42K, and 27K mol wt in AtT-20 cells. The relative proportions of the protein bands were unaltered by the use of whole cells or cell membranes or by the inclusion of dithiothreitol or antiproteolytic agents. With both [125I-Tyr11]S-14 and [125I-LTT]S-28, the 57K protein constituted the major labeled component, representing 70-75% of the total cross-linked proteins. Labeling of the three protein species by [125I-Tyr11]S-14 and [125I-LTT]S-28 was inhibited by both S-14 and S-28 in a dose-dependent manner. S-28 was 10-20 times more potent than S-14 for inhibiting the labeling by both ligands of the principal receptor species of 57K. By contrast, when a radioiodinated derivative of the octapeptide analog octreotide ([125I-Tyr3]SMS) was used as ligand, the 27K protein was preferentially labeled, whereas the 57K and 42K bands were detected only as minor components. Labeling of GH3 cells with [125I-Tyr11]S-14 and [125I-LTT]S-28 revealed three cross-linked proteins of 57K, 42K, and 27K mol wt similar to those observed in AtT-20 cells. However, in this cell line the 27K protein, not the 57K species, was the dominant component identified with these two ligands, comprising 40-50% of the total cross-linked proteins. These results suggest that there are three somatostatin receptor proteins of 57K, 42K, and 27K in pituitary cells. In AtT-20 cells, the 57K protein constitutes the major receptor protein labeled by [125I-Tyr11]S-14 and [125I-LTT]S-28, whereas the 27K protein is the major species labeled by [125I-Tyr3]SMS. The 27K, not the 57K, moiety is the principal receptor form in GH3 cells. Such ligand- and tissue-selective binding by the somatostatin receptor provides strong evidence for receptor molecular heterogeneity.

Multiple hormone storage by cells of the human pituitary.

While immunostaining serial semi-thin sections of acrylic resin-embedded normal human pituitary using antisera to human pituitary hormones, it became clear that several cells were stained by more than one antiserum. The tissue had been surgically excised from a patient with a prolactinoma. The tumor, which was immunoreactive only with antiprolactin antiserum, was distinctly different from the pieces of tissue under study which had normal pituitary architecture and demonstrated immunoreactivity with antisera against all six of the common pituitary hormones. A major immunoelectron microscopic investigation, using immunocolloidal gold and immunoperoxidase methods, revealed cells in which follicle-stimulating hormone (FSH), luteinizing hormone (LH), and prolactin (PRL) were co-localized to the same electron-dense granules. Some similar cells also possessed electron-lucent granules immunoreactive only for anti-PRL antiserum. Adrenocorticotrophic hormone (ACTH) and PRL were also found in the same cell but were very largely localized to separate, morphologically different populations of electron-dense and -lucent storage granules. By employing double immunolabeling, a few granules in the ACTH/PRL cells were shown to be immunoreactive to both anti-ACTH and anti-PRL antisera. The possibility that the multipotential stem cells is discussed.

Identification of thyrotroph-specific factors and cis-acting sequences of the murine thyrotropin beta subunit gene.

Pituitary thyrotroph cells specialize in the synthesis of TSH, and thus represent a model to study cell-specific gene expression. We have used the murine TSH beta (mTSH beta) gene promoter and TSH-producing and nonproducing transplantable tumors derived from murine thyrotroph cells, referred to as TtT-97 and MGH 101A, respectively, to identify nuclear factors which selectively interact with the mTSH beta gene. DNase I protection analyses demonstrate that factors present in TtT-97 nuclear extracts bind with high affinity to five separate sites in the TSH beta promoter region, denoted as distal D1 (-253 to -227) and proximal, P1 (-76 to -68), P2 (-106 to -98), P3 (-126 to -112), and P4 (-142 to -131) footprints. By contrast, non-TSH beta expressing thyrotroph cell nuclear extracts and L-cell nonpituitary cell extracts did not appear to footprint the D1 site; whereas the nonpituitary nuclear extracts revealed minimal DNase I protection in the P1-P4 regions. These data show that the distal D1 site is thyrotroph specific and contains a 6 base pair direct repeat sequence (5'-AGATAT-3'). Factor occupancy of the D1 site is protein dependent, occurs rapidly (less than 15 sec), is destabilized by 170 mM KCl, and results in an associated DNase I hypersensitive region. A double-stranded oligonucleotide spanning the D1 footprint competes only the distal factor binding region. Transfection of plasmid constructs containing progressive 5'-deletions of the mTSH beta promoter linked to the reporter gene luciferase into primary TtT-97 cells demonstrate a marked decrease in activity between the regions -270 and -79, which contains the D1 region.(ABSTRACT TRUNCATED AT 250 WORDS)

Cushing's disease in two sisters.

Hereditary Cushing's syndrome is an uncommon clinical entity, and most reported cases have been described in families with nodular adrenocortical dysplasia. Isolated cases of Cushing's disease (pituitary-dependent bilateral adrenal hyperplasia) have been reported in association with the multiple endocrine neoplasia syndrome, Type I (MEN I), but there are no published reports of pedigrees with more than one affected family member. Within a period of 8 months, two sisters presented with clinical findings suggestive of hypercortisolism, and Cushing's disease was confirmed by appropriate diagnostic studies. There was no evidence of any other endocrine excess syndrome in either patient. Transsphenoidal pituitary surgery confirmed the presence of an ACTH-immunostaining pituitary adenoma in each woman. The authors think this is the first report in the English literature of Cushing's disease in first-degree relatives.

Argyrophil organizer region proteins (AgNORs) in adenohypophysial cells and adenomas of the human pituitary.

Nucleolar organizer regions (NORs), claimed to represent valuable markers of tumor proliferation, were visualized on paraffin sections by an argyrophil method and counted in 18 nontumorous adenohypophyses and 132 pituitary tumors. The AgNOR counts could not be correlated with the hormonal activity of nontumorous and adenoma cells. In pituitary adenomas, the mean AgNOR values were higher than in their corresponding nontumorous cell types. Some adenomas, especially growth hormone (GH), and prolactin (PRL) cell adenomas, however, had AgNOR readings in the range of nontumorous cells. Long-acting somatostatin analog and bromocriptine treatment decreased AgNOR counts in GR- and PRL-producing tumors. Most, but not all invasive and/or recurrent adenomas had high AgNOR counts. In a corticotroph carcinoma, AgNORs were not higher than in the adenomas. These inconsistent results limit, at the present time, the use of AgNORs as reliable markers of cell proliferation in pituitary tumors. Further studies may help to establish the value of this promising method in pituitary pathology.

Chromogranin A and B messenger ribonucleic acids in pituitary and other normal and neoplastic human endocrine tissues.

The distribution of the messenger ribonucleic acids (mRNAs) for chromogranin A and B was analyzed by in situ hybridization in normal and neoplastic endocrine tissues using frozen and paraffin tissue sections. Combined in situ hybridization and immunochemical staining was also done on tissue sections from the same cases using a monoclonal antibody against chromogranin A (LK2H10). Most endocrine tumors expressed chromogranin A and B mRNAs as well as chromogranin A protein. Normal pituitary expressed chromogranin A and B mRNAs and chromogranin A protein in the anterior pituitary gland. Most of these cells were gonadotropic hormone-producing cells. Prolactinomas (5/5) did not express chromogranin A mRNA or protein, but contained chromogranin B mRNA. Null cell or nonfunctional adenomas (8/8) expressed chromogranin A and B mRNAs and reacted with antibody LK2H10. In some tumors such as Merkel cell carcinomas, insulinomas, and parathyroid adenomas, a stronger signal for chromogranin A mRNA was detected than for the immunoreactive proteins. These results indicate that in situ hybridization complements immunochemical techniques in the analysis of endocrine cells and neoplasms. The gene products for chromogranin A and B are widely distributed in many endocrine cells and tumors, but some neoplasms such as prolactinomas have a differential distribution of chromogranin A and B mRNA and proteins.

Peptides related to alpha-melanocyte-stimulating hormone are commonly produced by human pituitary corticotroph adenomas: no relationship with pars intermedia origin.

The presence of immunoreactive (ir)-alpha-MSH has been investigated by immunocytochemistry in 24 pituitary adenomas and one case of corticotroph hyperplasia causing Cushing's disease, in four adenomas causing Nelson's syndrome, and in ten 'silent' corticotroph adenomas. It was found that a high proportion of these adenomas have a population of cells containing ir-alpha-MSH in addition to ir-ACTH. In some instances, these adenomas were clearly not associated with the residual intermediate lobe of the pituitary. Radioimmunoassay of plasma from patients with Cushing's disease or Nelson's syndrome showed elevated levels of ir-alpha-MSH in the majority of cases. Characterization of the ir-alpha-MSH in adenoma cells by immunocytochemistry, using an antiserum selective for acetylated forms of alpha-MSH, suggested that only the desacetyl form was present in each case examined. High-performance liquid chromatography of adenoma tissue extracts revealed material co-eluting with acetylated forms of alpha-MSH in only one of six cases. These results have been compared with corticotroph adenomas in animal pituitary glands, and it is concluded that the presence of alpha-MSH peptides cannot be used as a marker for intermediate lobe tumours, and that desacetyl alpha-MSH is commonly produced by corticotroph adenomas.

[Immunohistochemical classification of hypophyseal adenomas]. / Imunohistochemická klasifikace adenomu hypofýzy.

Immunohistochemical analysis showed production of hormones in more than half of 88 pituitary adenomas that were classified mostly as chromophobe in basic staining during 3 years. Prolactin was the most often detected. A combined production of hormones especially of prolactin and growth hormone was not exceptional. Using a sensitive double indirect immunoperoxidase technique enables economy in spending primary antibodies. Evaluation of hormonal production is important for selection of therapy, as well as for settlement of prognosis.

Effects of estradiol on prolactin and growth hormone messenger RNAs in cultured normal and neoplastic (MtT/W15 and GH3) rat pituitary cells.

The effects of 17 beta-estradiol (E17 beta) on prolactin (PRL) cell proliferation and on the expression of PRL and growth hormone (GH) proteins and mRNAs were analyzed in cultured pituitary cells by immunocytochemistry, in situ hybridization, and Northern blot hybridization studies. Three different cell cultures were used: (a) normal pituitary cells; (b) GH3 tumor cell line; and (c) MtT/W15, a transplantable PRL and GH-producing pituitary tumor. E17 beta (10(-7) M) caused a significant increase in PRL cell proliferation in normal pituitary [3.9 +/- 0.4 versus 7.7 +/- 0.9% (SEM) of immunostained PRL cells with thymidine incorporation] [P less than 0.01] but produced a significant decrease in PRL cell proliferation in MtT/W15 primary cell cultures [6.7 +/- 1.0 versus 3.7 +/- 0.8%] [P less than 0.05]. PRL mRNA was significantly increased in normal pituitary and in GH3 tumor cells by E17 beta treatment. There was a significant decrease in PRL mRNA and an increase in GH mRNA expression in cultured MtT/W15 tumor cells by immunocytochemistry and in situ hybridization analyses. The percentage of cells producing both PRL and GH or mammosomatotropic cells analyzed by two different techniques declined after one week in culture in normal pituitary cells and in cultured MtT/W15 tumor cells after E17 beta treatment. These results show that E17 beta has a direct stimulatory effect on normal pituitary and GH3 cells and a direct inhibitory effect on MtT/W15 tumor cells with respect to cell proliferation and PRL hormone and mRNA expression.