Immunocytochemical and ultrastructural characterization of mammosomatotrope-, growth hormone-, and prolactin-cells from the gilthead sea bream (Sparus aurata l., Teleostei): an ontogenic study

Growth hormone (GH), prolactin (PRL), and mammosomatotrope (MS) cells of gilthead sea bream, Sparus aurata, a teleost fish, were studied in specimens from hatching to 15 months (adults) using conventional electron microscopy and an immunogold method using anti-tilapia GH sera and anti-chum salmon PRL serum. MS cells, immunoreactive to both anti-GH sera and anti-PRL sera, had been first identified in fish in a previous study in newly hatched larvae and in older larvae and juvenile specimens of Sparus aurata by light microscopic immunocytochemistry. In the present work, MS cells reacted positively to immunogold label only in older larvae and juveniles and their secretory granules immunoreacted with both GH and PRL antisera or with only one of them. MS cells were ultrastructurally similar to the PRL cells, with which they coincided in time. This is the first report on the ultrastructural characterization of MS cells in fish. In adults, the secretory granules of GH cells (immunoreactive to anti-GH serum) were mainly round, of variable size, and had a homogeneous, highly electron-dense content. Irregularly shaped secretory granules were also present. PRL cells (immunoreactive to anti-PRL serum) were usually observed in a follicular arrangement; they showed few, small, and mainly round secretory granules with a homogeneous and high or medium electron-dense content. Some oval or elongated secretory granules were also observed. GH and PRL cells that showed involutive features were also found. In newly hatched larvae, GH, PRL, and MS cells could not be distinguished either by their ultrastructure or by the immunogold labeling of the secretory granules. In 1-day-old larvae, presumptive GH and PRL cells were observed according to their position in the pituitary gland. In 2-day-old larvae, a few cells showed some of the ultrastructural features described for GH and PRL cells of adults. During development, the number, size, and shape of the secretory granules in both cell types clearly increased and the organelles developed gradually. Some GH cells were found undergoing mitosis.     

Sorting of three secretory proteins to distinct secretory granules in acidophilic cells of cow anterior pituitary

The distribution of three proteins discharged by regulated exocytosis--growth hormone (GH), prolactin (PRL), and secretogranin II (SgII)--was investigated by double immunolabeling of ultrathin frozen sections in the acidophilic cells of the bovine pituitary. In mammotrophs, heavy PRL labeling was observed over secretory granule matrices (including the immature matrices at the trans Golgi surface) and also over Golgi cisternae. In contrast, in somatotrophs heavy GH labeling was restricted to the granule matrices; vesicles and tubules at the trans Golgi region showed some and the Golgi cisternae only sparse labeling. All somatotrophs and mammotrophs were heavily positive for GH and PRL, respectively, and were found to contain small amounts of the other hormone as well, which, however, was almost completely absent from granules, and was more concentrated in the Golgi complex, admixed with the predominant hormone. Mixed somatomammotrophs (approximately 26% of the acidophilic cells) were heavily positive for both GH and PRL. Although admixed within Golgi cisternae, the two hormones were stored separately within distinct granule types. A third type of granule was found to contain SgII. Spillage of small amounts of each of the three secretory proteins into granules containing predominantly another protein was common, but true intermixing (i.e., coexistence within single granules of comparable amounts of two proteins) was very rare. It is concluded that in the regulated pathway of acidophilic pituitary, cell mechanisms exist that cause sorting of the three secretory proteins investigated. Such mechanisms operate beyond the Golgi cisternae, possibly at the sites where condensation of secretion products into granule matrices takes place.     

Two types of mammosomatotropes in mouse adenohypophysis

Summary Two types of mammosomatotropes (MS), the small-granule and vesicle-granule MS, were detected in mouse adenohypophysis by electron microscopy and immunohistochemistry. Both cell-types were immunoreactive to prolactin (PRL) and growth hormone (GH) antisera. The small-granule MS contained small, round, solid secretory granules about 100 nm in diameter, and were smaller than the classical GH and PRL cell-types. The vesicle-granule MS contained secretory granules like cored vesicles, and were larger than classical GH and PRL cells. Small-granule MS were immunoreactive to both PRL and GH antisera in the same region of the cell cytoplasm; the vesicle-granule MS, however, were immunoreactive to only PRL antiserum in most cytoplasmic areas, and a positive response to both PRL and GH antisera was confined to only certain small areas.     

Regulation of chromogranin B/secretogranin I and secretogranin II storage in GH4C1 cells

GH4C1 cells are a rat pituitary tumor cell strain in which the level of cellular prolactin (PRL) and PRL-containing secretory granules can be regulated by hormone treatment. The chromogranins/secretogranins (Sg) are a family of secretory proteins which are widely distributed in the secretory granules of endocrine and neuronal cells. In the present study, we investigated in GH4C1 cell cultures the regulation of the cell content of the Sg by immunoblotting and the relationship between the storage of Sg I and Sg II and PRL by double immunocytochemistry. GH4C1 cells grown in the presence of gelded horse serum, a condition in which these cells contain a low level of secretory granules, contained low levels of PRL, Sg I, and Sg II. Treatment of GH4C1 cells with a combination of 17 beta-estradiol, insulin, and epidermal growth factor for 3 days, known to induce a marked increase in the number of secretory granules, increased the cell contents of PRL, Sg I, and Sg II. To determine whether the induction of PRL was morphologically associated with that of the Sg, the distribution of PRL and the Sg was determined by double immunofluorescence microscopy. After hormone treatment, 54% of cells showed positive PRL immunoreactivity, fluorescence being extranuclear and consistent with staining of the Golgi zone and secretory granules. Forty-six percent of PRL-positive cells stained coincidently for Sg I, while 72% of the PRL cells were also reactive with anti-Sg II. To determine whether PRL storage was associated with storage of at least one of the Sg, cells were stained with anti-PRL and anti-Sg I and anti-Sg II together. Eighty-six percent of PRL cells stained for one or the other of the Sg. Therefore, PRL storage in GH4C1 cell cultures is closely but not completely associated with the storage of Sg I and/or II.     

Architecture of the two metal-binding sites in prolactin

Metal binding by members of the growth hormone (GH) family of hematopoietic cytokines has been a subject of considerable interest. However, beyond appreciation of its role in reversible packing of GH proteins in secretory granules, the molecular mechanisms of metal binding and granule formation remain poorly understood. Here, we investigate metal binding by a GH family member prolactin (PRL) using paramagnetic metal titration and chelation experiments. Cu²⁺-mediated paramagnetic relaxation enhancement measurements identified two partial metal-binding sites on the opposite faces of PRL composed of residues H30/H180 and E93/H97, respectively. Coordination of metal ions by these two sites causes formation of inter-molecular bridges between the PRL protomers and enables formation of reversible higher aggregates. These findings in vitro suggest a model for reversible packaging of PRL in secretory granules. The proposed mechanism of metal-promoted PRL aggregation lends insight and support to the previously suggested role of metal coordination in secretory granule formation by GH proteins.     

Electron microscopic immunocytochemistry of prolactin secreting cells (PRL-cells) and growth hormone secreting cells (GH-cells) in the anterior pituitary gland of partially hepatectomized rats

After 60% hepatectomy in rats, prolactin secreting cells (PRL-cells) and growth hormone secreting cells (GH-cells) of the anterior pituitary gland were distinctly identified by the protein A-gold procedure combined with electron microscopy. The animals were sacrificed by the decapitation at midnight at intervals of about 28, 76 and 124 h after the operation. The principal changes can be summarized as follows; (1) Hypertrophy of the Golgi complex, (2) Dilation of the rough endoplasmic reticulum (RER), (3) Increased numbers of secretory granules (markedly in GH-cells), and occurrence of granule extrusion or exocytosis, (4) Increased numbers of lysosomes, which were mostly seen at 124 h after the operation. These ultrastructural changes were remarkably observed in both PRL and GH-cells. Especially, the noticeable changes in PRL-cells after partial hepatectomy in the rat were new findings in this study. The above results suggest that hepatectomy induced synthesis and release of GH and PRL in cells from pars distalis.     

Immunocytochemical identification of growth hormone cells in the adenohypophysis of the golden hamster

Somatotrophs or growth hormone (GH) cells in the adenohypophysis of golden hamsters were identified by immunocytochemical staining with polyclonal rabbit anti-human GH. They were oval or columnar in shape, and had secretory granules of two size ranges, 90-150 nm and 280-320 nm, which were present in the same cells; no subtypes of GH cells were observed. Secretory granules were located in the peripheral portion of the cytoplasm or concentrated at the vascular pole of the cell. Flattened cisternae of the rough endoplasmic reticulum in parallel array and a moderately developed Golgi apparatus were often found in the cytoplasm. No sex difference was noticed in the population ratio of GH cells. Immunocytochemical staining with anti-GH or anti-prolactin (PRL) antibodies on separate adjacent sections revealed that the GH and PRL were stored in two different cell types.     

Induction of mammotroph development by a combination of epidermal growth factor,insulin, and estradiol-17beta in rat pituitary tumor GH3 cells

Several reports have indicated that prolactin-secreting cells (PRL cells) are generated from growth hormone-secreting cells (GH cells). We have shown that treatment with a combination of epidermal growth factor (EGF), insulin, and estradiol-17beta (E (2)) induces the appearance of PRL cells in pituitary tumor GH3 cells. The aim of the present study was to clarify the involvement of mitosis in the cytogenesis of PRL cells in rat pituitary and GH3 cells. The effects of the treatment with EGF, insulin and E(2) on DNA-replication were studied by detecting the uptake of bromodeoxyuridine (BrdU) into the nucleus. In cultured rat pituitary cells, BrdU-labeled PRL cells were observed irrespective of the hormone treatment. In GH3 cells, BrdU-labeled GH cells and mammosomatotrophs (MS cells) were detected; BrdU-labeled PRL cells were not detected, however, when GH3 cells were treated with BrdU for 3 hr and then immediately examined for BrdU-labeling. BrdU-labeled PRL cells were found only when GH3 cells treated with BrdU were allowed to grow for another 3 days. This finding suggests that during the additional 3-day culture, BrdU-labeled PRL cells were generated from BrdU-labeled cells other than PRL cells. These results indicate that PRL cells are transdifferentiated from GH cells or MS cells in GH3 cells by a combined treatment with EGF, insulin and E(2), while PRL cells in rat pituitaries are able to proliferate in response to the hormone treatment. Thus, there may be two pathways for cytogenesis of PRL cells: the transdifferentiation of GH cells or MS cells, and a self-duplication of PRL cells.     

Subcellular distribution of secretogranins I and II in GH3 rat tumoral prolactin (PRL) cells as revealed by electron microscopic immunocytochemistry

The GH3 rat pituitary cell line which secretes prolactin (PRL) is characterized by the paucity and small size of secretory granules. We looked for the presence, in these cells and in normal PRL cells, of two acidic tyrosine-sulfated proteins which are widely distributed in dense-core secretory granules of endocrine and neuronal cells, secretogranins I and II, using immunofluorescence and electron microscope immunoperoxidase techniques. Both secretogranins were detected in secretory granules of GH3 cells and of normal cells. Moreover, with our pre-embedding approach, secretogranins were localized within some RER cisternae and within all sacules of the Golgi stacks in both PRL cell models. A few small vesicles, large dilated vacuolar or multivesicular structures, and some lysosome-like structures were also immunoreactive. Double localization of secretogranins and PRL performed on GH3 cells by immunofluorescence indicated that all cells contained secretogranins I and II, whereas only 50-70% of the cells contained PRL. Moreover, in the case of hormone treatment known to increase the number of secretory granules, most if not all mature secretory granules were immunoreactive for secretogranins, whereas in certain cells some of the granules were apparently not immunoreactive for PRL. These immunocytochemical observations show that GH3 cells, which under normal conditions form only a small number of secretory granules, produce secretogranins and package them into these granules.     

Growth hormone and prolactin immunoreactivity in the pituitary gland of postnatal little (lit) mice

Homozygous little (lit/lit) mutant mice exhibit a growth lag which is manifested at approximately two weeks postnatally. Functional aspects of the development of pituitary growth hormone (GH) cells and prolactin (PRL) cells were thus analyzed by means of colloidal gold immunocytochemistry at the ultrastructural level in lit/lit mice and their normal counterparts ranging in age from 5 days postnatally to adulthood. In the adult normal and lit/lit pituitaries, secretory granules in GH cells and PRL cells showed a positive immunoreaction to their respective antisera, as did granules in both cell-types at 5 days postnatally. By 14 days some GH cells in lit/lit pituitaries appeared to be less densely populated with granules than GH cells in normal pituitaries, but a positive immunoreaction continued to occur even in sparsely granulated GH cells. PRL cells showed ultrastructural features in lit/lit pituitaries which were similar to those in normal mice, and immunoreactivity was present at all stages examined. The results indicate that since differences in granule reactivity were not evident between lit/lit and normal GH cells, despite ultrastructural morphologic differences which were present by 14 days postnatally, manifestations of the defect in lit/lit may be primarily quantitative in terms of numbers of granules and/or numbers of GH cells. With respect to PRL cells, neither morphologic nor functional aberrations could be observed; thus, a deficit in PRL hormone production might be the result of a more subtle defect than that in GH cells.     

Effect of vasoactive intestinal polypeptide (VIP) on growth hormone (GH) and prolactin (PRL) release and cell morphology in human pituitary adenoma cell cultures

Six GH adenomas and three prolactinomas were investigated by light- and electron-microscopic morphological and immunocytochemical methods and the effect of vasoactive intestinal polypeptide (VIP) on growth hormone (GH) and prolactin (PRL) secretion was tested in vitro. The tumour cells of the acromegalic patients revealed both GH and PRL immunoreactivity while prolactinomas showed only PRL activity. All the adenomas stained immunocytochemically also for VIP. By electron microscopy, the tumours included two densely and two sparsely granulated GH, two mixed GH/PRL, and three sparsely granulated PRL adenomas. The dissociated cells were explanted, and cultured in vitro. The cultures in micro test plates were treated with VIP at different concentrations between 10(-5)-10(-12) M. GH and PRL contents in the culture media were measured by radioimmunoassay. GH release was significantly stimulated by VIP in a dose-dependent manner over the whole concentration range, while VIP was effective on the PRL release only at 10(-6)-10(-7) M concentration. The cells of a mixed adenoma were grown in Petri dishes and used for ultrastructural and immunocytochemical studies. The cytoplasmic structure of the cells treated with VIP corresponded to that of active hormone-secreting cells with large ergastoplasmic fields and Golgi zones containing secretory granules. Massive exocytotic events were encountered mainly in the GH-type cells. GH and PRL double immunocytochemistry showed the predominance of GH cells, many of them containing low amounts of PRL as well. Cells predominantly containing PRL were spread among them, they also might contain GH as well. Some of the cells contained only a single immunoreactive hormone. The intensity of gold labelling of the secretory granules appeared higher in the VIP-treated cells than in the untreated control ones which showed a cytoplasmic structure characteristic of glandular cells with low secretory activity. As all the adenoma cells both contained and reacted to VIP, our results are in agreement with an autocrine or paracrine effect of this peptide. The fine structure of the cells in the cultures treated with VIP supply an additional argument to the assumption that VIP may serve as a growth factor for these cell types.     

Conversion of growth hormone-secreting cells into prolactin-secreting cells and its promotion by insulin and insulin-like growth factor-1 in vitro

The newly established rat pituitary cell line, MtT/S, has pituitary somatotroph (growth hormone-producing cell)-like characteristics, i.e., the cells produce growth hormone (GH), possess GH-immunopositive secretory granules, and respond to GH-releasing hormone. When MtT/S cells were cultured in regular medium no prolactin (PRL) cells were observed and PRL was not detected, by radioimmunoassay or Western blot analysis, in the medium or the cells. However, GH production and the GH cell population decreased markedly when the cells were incubated with insulin or insulin-like growth factor-1 (IGF-1). After stimulation with insulin or IGF-1 there was a 2-day lag period, then some PRL was detected in the medium; after 5 days a number of PRL cells appeared. Double immunocytochemistry indicated clearly that no cell contained both PRL and GH. These results show that insulin and IGF-1 stimulate conversion of MtT/S cell line GH cells to PRL cells. This suggests that the MtT/S cell line is an excellent model system which shows the GH-cell/PRL-cell lineage.     

Immunohistochemical and ultrastructural study of anterior pituitary cells in the female Afghan pika,Ochotona rufescens rufescens

An immunohistochemical study of the anterior pituitary gland of the female Afghan pika was carried out to distinguish the ultrastructural features of GH, PRL, ACTH, TSH and LH cells. The histochemically identified GH cells resembled ultrastructurally oval or round GH cells of the rat laden with large, dense secretory granules. PRL cells were divided into three subtypes based on differences in the diameter of their spherical secretory granules. They lacked polymorphic or irregularly shaped secretory granules. ACTH cells resembled ultrastructurally, in some respects, Siperstein's "corticotrophs" of the rat with peripheral arrangement of secretory granules. However, they were not always stellate, but elongate or angular in shape. The dense secretory granules were concentrated in the peripheral area of cytoplasm. TSH cells were non-stellate, but usually oval in shape, containing the smallest spherical secretory granules (100-200 nm in diameter). Almost all LH cells reacted also with FSH antiserum. They were irregular in shape, sometimes in contact with or surrounded the GH cells. They contained an abundance of medium-sized secretory granules (140-260 nm in diameter) which were larger than those in the LH cells of the female rat throughout the estrous cycle. Large secretory granules in the LH cells of the female pika seemed to be related to the endocrine state of persistent estrus.     

Calcium and the secretory cycle of prolactin cells: a cytochemical and ultrastructural study of dopamine inhibition and monobutyryl cyclic AMP-stimulation of prolactin secretion

To identify intracellular calcium pools that may be involved in the secretory process in prolactin (PRL) cells, hemi pituitaries were incubated in medium containing 10(-6) M dopamine, 5 mM cyclic cAMP (experimentals), or in medium alone (controls) and then processed for electron microscopy using potassium pyroantimonate to localize intracellular calcium. PRL in the medium was measured by radioimmunoassay. The concentration of antimonate associated with mitochondria, Golgi saccules, and secretory granules was estimated. Dopamine inhibition of PRL secretion (> 80% at 1, 2, 3 h) resulted in accumulation of secretory granules in all stages of maturation and dilation of Golgi saccules at 2 and 3 h, accompanied by increased mitochondria antimonate and increased Golgi-associated antimonate. Cyclic AMP stimulation of secretion (635% at 5 min., declining to 34% at 1 h) resulted in marked exocytosis at 5 and 15 min., declining after 30 min. Mitochondrial antimonate decreased after 30 min. Stimulated cells exhibited numerous coated membrane structures at or near exocytotic pits and an amassing of microvesicles at the margin of the Golgi apparatus. Although some secretory granules consistently exhibited reactivity to antimonate (unchanged by inhibition or stimulation), plasma membrane, and granule membrane translocated to the plasma membrane during exocytosis, were not reactive.     

Immunocytochemical study of the GH cells in the anterior pituitary gland of human fetus II. Anencephalic fetus

In order to elucidate the effects of hypothalamic regulation on the morphology of GH cells, light and electron microscopic immunocytochemical examinations were carried out comparing GH cells in the anterior pituitary gland of anencephalic fetus with those of normal fetuses. Three types of GH cells were identified in the anterior pituitary gland of anencephalic fetus as well as in the normal fetus. Type-I is a small, round cell containing a few small secretory granules. Type-III is a large, polygonal cell with numerous large secretory granules. Type-II is a polygonal cell with medium-sized secretory granules. The Type-II GH cell was predominant in both anencephalic and normal fetuses. The most striking difference between anencephalic and normal fetuses was the presence of atypical forms of the Type II cell. These were polygonal cells containing secretory granules, which were either immunopositive or immunonegative to anti-human GH (anti-hGH) serum. Furthermore, two other types of GH cells were identified. The somatomammotroph (SM cell) contained GH and PRL in different granules within the same cell. Also, a different type of the GH cell was noted containing two varieties of secretory granules; one was immunolabeled only with anti-hGH and the other was not immunolabeled to either anti-hGH or anti-human PRL (anti-hPRL). From these results, we suggest that an absence of hypothalamic regulation in the anencehpalic does not seriously modify GH cell morphology but induces an altered GH storage pattern in some of the cells.     

Isolated prolactin and growth hormone granules from bovine pituitary: content and stability are seasonally dependent

Secretory granules containing prolactin (PRL) and growth hormone (GH) as essentially the only proteins were isolated by centrifugation. PRL and GH varied reciprocally in the granule preparations with the seasons. During winter PRL content was lowest (20%) and GH highest (80%); during summer the converse obtained: PRL, 70% and GH,, 30%. Both hormones were in almost equal proportion during the spring. The amount of either hormone released from granules and pituitary slices was directly related to its relative content in the gland. The pattern of PRL release from secretory granules and pituitary tissue in vitro was similar to that reported for blood levels in ruminants: low during winter and high during summer. It is concluded that seasonal factors affect primarily the synthesis and/or storage of PRL and GH, and there exists a direct relationship between intracellular stores and release.     

Distribution of growth hormone-like cells in the pituitary of adult sea lampreys,Petromyzon marinus

Growth hormone (GH), prolactin (PRL) and somatolactin (SL) are members of a pituitary hormone family that are believed to have evolved from a common ancestral gene by duplication and subsequent divergence. Since these hormones are found both in bony fish and cartilaginous fish, their ancestral form(s) should be present in the Agnatha. Thus, although there is no convincing evidence that the lamprey pituitary secretes GH or PRL, GH- and/or PRL-like immunoreactivity was examined in the pituitary of adult sea lampreys (Petromyzon marinus), using antibodies to GHs, PRLs and SL of mammalian and/or fish origins. Our initial attempt with ordinary immunohistochemical procedures failed to detect any positive reactions in the lamprey pituitary. Following the hydrated autoclave pretreatment of the sections, anti-salmon GH, anti-salmon PRL and anti-blue shark GH gave positive reactions in most cells distributed in the dorsal half of the proximal pars distalis. These results suggest that the material immunoreactive to those antibodies is related, to some extent, to GH/PRL, but enhancement of immunoreactivity to reveal this by the hydrated autoclave pretreatment of sections is needed due to low crossreactivity. The similarity of the topographic distributions within the pituitary between lampreys and teleosts suggests that lamprey GH/PRL-like cells are GH cells of the lamprey.     

Mammosomatotrophs develop within mammotroph clusters in bovine adenohypophysis

The existence and distribution of mammosomatotrophs (MS cells) containing growth hormone (GH) and prolactin (PRL) in bovine adenohypophysis were detailed by a combined method of mirror sections and immunohistochemical staining. MS cells always occurred in bovine adenohypophysis but their number was low. In the midsagittal plane, the cells were observed in the hind dorsal, hind ventral and fore ventral region abundant in GH and PRL cells. Whereas, in the zona tuberalis where GH and PRL cells were less frequent, MS cells were not detected. MS cells were invariably solitarily distributed within mammotroph (PRL cell) clusters but not within somatotroph (GH cell) clusters. The proportion of MS cells declined as the ages proceeded and the appearance was spatially related to the arrangement of PRL cells. These findings indicated that, in bovine adenohypophysis, MS cells were differentially distributed and occurred in PRL cell clusters. The results strongly suggest that MS cells originate in GH cells pre-existed within PRL cell clusters with special reference to the functional activation of PRL cells.     

Intracellular Routing and Release of Caseins and Growth Hormone Produced into Milk from Transgenic Mice

Growth hormone (GH) secretion, in mammary tissue from transgenic mice, containing a chimeric gene composed of the regulatory region of whey acidic protein gene and the structural region of GH gene, was compared to casein secretion. GH was expressed in milk and for a small percentage (1:1000) in blood as revealed by SDS-polyacrylamide gel electrophoresis and radio-immunoassay. As attested by immunofluorescence and immunogold electron microscopy, caseins and GH followed the same secretory pathway. However, contrary to caseins, which are essentially in micellar form, GH was detected in a nonaggregated form in secretory vesicles and in the lumen of the acini. Newly synthesized caseins and GH were carried simultaneously, mainly to the lumen of the acini, but also to the base of the cell. Secretion of newly synthesized proteins was increased by prolactin (PRL). As shown by immunoblotting, the proportion of GH versus other proteins, secreted in the presence of PRL was not modified, suggesting that GH secretion is subjected to the same hormonal regulation by PRL as other milk proteins. These results show that, in lactating mammary epithelial cells from transgenic mice, a recombinant GH and the caseins are carried simultaneously to the lumen and suggest that secretion of both proteins is increased by PRL during the same time course. Transport of these newly synthesized proteins occurs also to the base of the cell.     

Epidermal growth factor, insulin, and estrogen stimulate development of prolactin-secreting cells in cultures of GH3 cells

Pituitary tumor GH3 cells synthesize and secrete both growth hormone (GH) and prolactin (PRL). Morphological and functional changes of GH3 cells induced by epidermal growth factor (EGF, 10 nM), insulin (300 nM), and estradiol-17beta (E2, 1 nM) were studied. Treatment of cultures of GH3 cells for 6 days with EGF, insulin, or E2 alone, and with EGF plus E2 did not affect the total number of GH3 cells, but a combination of EGF, insulin, and E2 decreased the total number of GH3 cells compared with control treatment. DNA-synthesizing cells were detected by monitoring 5-bromo-2'-deoxyuridine (BrdU) uptake. EGF, E2, or a combination of EGF, insulin, and E2 significantly decreased the proportion of BrdU-labeled cells (21.1+/-1.7%, 21.0+/-1.4%, 18.2+/-1.3%; P<0.05, P<0.05, P<0.01, respectively) compared with control treatment (28.6+/-1.5%), but insulin did not (31.4+/-2.4%). Immunocytochemical analysis of GH3 cells cultured in 5% fetal calf serum-supplemented medium (control) showed that about 70% of all GH3 cells were GH-immunoreactive cells (GH-ir cells), apparently containing only GH, and 14% were mammosomatotrophs (MS cells), containing both GH and PRL, while PRL-immunoreactive cells (PRL-ir cells), containing only PRL, were not detected. No GH or PRL immunoreactivity could be detected in the remaining cells (15%). EGF decreased the proportion of GH-ir cells. The effects of EGF were enhanced by simultaneous exposure to insulin and E2; this decreased the proportion of GH-ir cells to about 20% of the total GH3 cells and significantly increased the proportion of MS cells to 300% of controls. Treatment with EGF plus insulin, EGF plus E2, or a combination of EGF, insulin, and E2 all stimulated the appearance of PRL-ir cells. Exposure to EGF caused a significant decrease in GH mRNA (P<0.01) and a significant increase in PRL mRNA (P<0.05). These observations suggest that EGF is closely involved in differentiation of PRL-ir cells from GH-ir cells via MS cells in GH3 cell cultures. Cytosine arabinoside (10(-7) M), an inhibitor of cell division, did not affect the changes in proportion of the three cell types induced by treatment with a combination of EGF, insulin, and E2. It is therefore probable that the transdifferentiation does not require mitosis of the GH3 cells.