Galanin within the normal and hyperplastic anterior pituitary gland: localization, secretion, and functional analysis in normal and human growth hormone-releasing hormone transgenic mice.

Studies evaluating estrogen-induced anterior pituitary tumors revealed a strong direct correlation between expression of the peptide galanin and tumor growth. To evaluate further the potential roles of galanin in the hyperplastic pituitary, we used a model of estrogen-independent anterior pituitary tumor formation, the male human GH-releasing hormone (hGHRH) transgenic mouse. Pituitaries of hGHRH transgenic mice are characterized by a hyperplasia of somatotrophs and contain markedly elevated levels of galanin. We examined the population of galanin-producing pituitary cells in 4- to 6-month-old male hGHRH transgenic mice and their nontransgenic siblings. The percentage of galanin-containing pituitary cells was significantly increased within the anterior pituitaries of hGHRH transgenic mice. By using the cell immunoblot assay we found that the basal secretion of galanin and GH from individual pituitary cells of hGHRH transgenic mice was significantly greater than that from pituitary cells of nontransgenic mice. By modifying the cell immunoblot assay, we determined that somatotrophs from both hGHRH transgenic and normal mice that were positive for galanin immunoreactivity secreted significantly greater amounts of GH than those somatotrophs devoid of galanin immunoreactivity. Moreover, immunoneutralization of galanin significantly decreased GH secretion from pituitary cells obtained from hGHRH transgenic mice. Thus, we now show that the increased levels of galanin peptide within the hyperplastic pituitaries of hGHRH transgenic mice are due to an increase in the population of cells containing galanin, and that galanin participates in the augmented secretion of GH from hyperplastic proliferating pituitary cells.

Immunohistochemical localization of neuropeptides and neurotransmitters in the nucleus solitarius.

The nucleus tractus solitarii (NTS), which receives visceral afferent information from the cardiovascular, respiratory, gastrointestinal and taste systems, contains multiple neurotransmitters and neuropeptides throughout its rostral to caudal extent. The neurotransmitters and neuropeptides immunoreactivity is located predominately in varicose fibers and small puncta throughout the neuropil. In addition, immunoreactive NTS neurons for a variety of neurotransmitters and neuropeptides are present in subnuclear regions. The neuroactive substances localized immunohistochemically in the NTS include acetylcholine, the neuropeptides, substance P, methionine- and leucine-enkephalin, beta-endorphin, cholecystokinin, neurotensin, galanin, calcitonin gene-related peptide, somatostatin, FMRMamide, neuropeptide Y, angiotensin II, vasoactive intestinal polypeptide, vasopressin, oxytocin, thyrotropin-releasing hormone, luteinizing hormone-releasing hormone, atrial natriuretic peptide, the catecholamines, dopamine, norepinephrine, epinephrine, serotonin, histamine and the amino acids, GABA and glutamate. The pattern of innervation for each neurotransmitter and neuropeptide is not homogeneously distributed throughout the NTS. Each substance has a unique pattern within the NTS as each subnuclear region contains different immunohistochemical staining patterns and densities of fibers. At the ultrastructural level both neurotransmitters and neuropeptides are present in synaptic terminals that are in contact with different parts of the neuronal membranes. Typically, the labeled terminals contain both small, clear vesicles and large, dense core vesicles with the exception of synaptic terminals containing acetylcholine, GABA and glutamate which do not typically have the large, dense core vesicles. The most frequent post-synaptic target are dendrites and spinous processes. Less frequently, synaptic contacts are present on the cell soma.

Vasoactive intestinal polypeptide mRNA and peptide levels are decreased in the anterior pituitary of the human growth hormone-releasing hormone transgenic mouse.

We recently reported that galanin gene expression is markedly increased in the hyperplastic anterior pituitary gland of the human growth hormone-releasing hormone (hGHRH) transgenic mouse. To determine if another pituitary peptide hormone with putative growth-promoting activity is similarly affected, or if this effect is specific to the peptide galanin, we examined vasoactive intestinal polypeptide (VIP) gene expression in the hypothalamic-pituitary axis of male hGHRH transgenic and non-transgenic mice. The objectives were to: 1) assess VIP peptide concentrations, 2) estimate relative differences in VIP mRNA levels, 3) determine the effects of acute treatment with 17beta-estradiol on VIP peptide and mRNA levels, and 4) quantify the density of immunoreactive VIP pituitary cells by immunohistochemistry. Four to five month old male hGHRH transgenic mice and their non-transgenic siblings were identified by PCR. Immunoreactive VIP concentrations were decreased by 50% in the anterior pituitary glands of hGHRH transgenic mice as compared to non-transgenic siblings. In contrast, no differences in immunoreactive VIP concentrations were observed in the hypothalamus or frontal cerebral cortex of transgenic and non-transgenic mice. Treatment with 17beta-estradiol significantly increased VIP concentrations in the anterior pituitary gland of both transgenic and non-transgenic mice; however, VIP peptide concentrations in the anterior pituitary glands of hGHRH transgenic mice remained 50% lower. Relative differences in VIP mRNA levels were estimated by RT-PCR, and were found to be 2.5-fold higher in the anterior pituitary glands of non-transgenic mice. In contrast, no differences in VIP mRNA levels in the cerebral cortex were detected between transgenic and non-transgenic mice. Treatment with 17beta-estradiol increased VIP mRNA levels in the anterior pituitary, but not in the cerebral cortex. In concert with the changes in VIP peptide and mRNA, the density of immunoreactive VIP pituitary cells was decreased approximately 50% in hGHRH transgenic mice. In conclusion, unlike galanin gene expression, VIP peptide and mRNA levels are significantly decreased in the anterior pituitary gland of hGHRH transgenic mice. Moreover, these changes appear to be tissue-specific and are likely due, in part, to the decrease in the density of VIP-containing pituitary cells in the hyperplastic pituitary. Although the pituitary cell type(s) synthesizing VIP remains unclear, these data suggest that VIP in the anterior pituitary is not stimulating pituitary tumor development in hGHRH transgenic mice.

Estrogen-receptive neurons in the anteroventral periventricular nucleus are synaptic targets of the suprachiasmatic nucleus and peri-suprachiasmatic region.

The anteroventral periventricular nucleus (AVPv) in the rat preoptic area is a key site underlying control of the steroid dependent preovulatory gonadotropin surge. Estrogen and progesterone receptor-containing neurons in the preoptic/hypothalamic continuum, particularly those in the AVPv, are believed to transduce steroidal signals and, in turn convey this information to the LHRH system, which lacks steroid receptors. In addition to the influence of the gonadal steroids, the precise timing of the preovulatory gonadotropin surge is believed to be regulated by the hypothalamic suprachiasmatic nucleus (SCN). The SCN and peri-SCN neurons send efferent projections rostrally to the anterior preoptic area suggesting that circadian signals are communicated synaptically to steroid-responsive neurons in the AVPv. To test this hypothesis, ultrastructural double label immunocytochemistry was conducted to determine whether SCN efferents contact estrogen receptor-immunoreactive neurons in the AVPv. Brain sections with SCN injections of phaseolus vulgaris leucoagglutinin (PHA-L) were immunostained for estrogen receptors and PHA-L. Light and electron microscopic data show that the anterior preoptic area received robust PHA-L-immunoreactive efferents from SCN neurons and immediately adjacent subparaventricular zone. In particular, the AVPv contained abundant labeled fibers and terminal boutons. Ultrastructurally, SCN- and subparaventricular zone-derived terminals synaptically contacted the perikaryon of many estrogen receptor-immunoreactive neurons in the AVPv. The perikarya of unlabeled neurons were also contacted, but the majority of the labeled contacts were observed upon neuronal processes. These results demonstrate that estrogen responsive AVPv neurons are regulated by SCN efferents. Furthermore, the present data provide strong support to the idea of collective control of pituitary gonadotropin release by steroid sensitive and circadian signal neural pathways.

Estrous cycle-associated axosomatic synaptic plasticity upon estrogen receptive neurons in the rat preoptic area.

This study examined the hypothesis that synaptic relationships change on a daily basis in the anterior preoptic area of the intact, cycling female rat. Specifically, the anteroventral periventricular nucleus (AVPv) was chosen for analysis due to its abundant estrogen- and progesterone-receptive neurons and its critical role in the control of gonadotropin secretion. Ultrastructural analysis revealed that in the 24 h between proestrus and estrus, there was a 39% increase in axosomatic synapses upon AVPv neurons. In the subsequent 24 h to metestrus, the number of synapses decreased by 22%. Additional data showed that ovariectomy resulted in more axosomatic synapses in the AVPv relative to proestrus (46%) and metestrus (35%). Another component of the study investigated synaptic relationships on estrogen receptor-immunoreactive (ER-ir) and non-ER-ir neurons across the estrous cycle. ER-ir neurons received significantly more synaptic input at proestrus and estrus than did non-ER-ir neurons. At metestrus and following ovariectomy, no significant differences were present. The present study indicates that estrous cycle-associated synaptic plasticity occurs in the AVPv and lends further support to the critical role of this nucleus in regulation of gonadotropin secretion. Moreover, it provides the first evidence that ER- and non-ER-ir neurons are differentially innervated.

Distribution of [3H]quinuclidinyl benzilate, [3H]nicotine, and [125I]alpha-bungarotoxin binding sites in the nucleus tractus solitarii of the cat.

The distribution of muscarinic and nicotinic cholinergic binding sites in the cat nucleus tractus solitarii was studied by the technique of in vitro autoradiography. Using the antagonist [3H]quinuclidinyl benzilate, muscarinic binding sites were differentially located in subdivisions of the nucleus tractus solitarii. The majority of muscarinic binding sites were located predominantly in the caudal half of the nucleus, reaching their greatest amounts at the mid levels of the nucleus tractus solitarii. The medial, dorsolateral, intermediate, and interstitial subdivisions contained the highest densities of quinuclidinyl benzilate binding sites. Nicotinic cholinergic binding sites, using [3H]nicotine and [125I]alpha-bungarotoxin, had unique patterns of distribution. With [3H]nicotine the majority of binding sites were located in rostral levels of the nucleus with very few binding sites present in the caudal half. In contrast, [125I]alpha-bungarotoxin binding sites were present mainly in subdivisions located in the caudal half of the nucleus, i.e., commissural, ventrolateral, dorsolateral, medial, and intermediate subdivisions, and dropped off precipitously at more rostral levels. The differential distribution of [3H]nicotine and [125I]alpha-bungarotoxin suggests the two ligands may be labeling different types of nicotinic binding sites in the nucleus tractus solitarii. The unique distribution of muscarinic and nicotinic cholinergic binding sites in the various subdivisions of the nucleus solitarii suggests that muscarinic and nicotine mechanisms may play an active role in the regulation of the diverse autonomic functions at the level of the nucleus tractus solitarii.

Colocalization of galanin and prolactin within secretory granules of anterior pituitary cells in estrogen-treated Fischer 344 rats.

Galanin is localized within specific cell types of the rat anterior pituitary gland (AP). Immunocytochemical studies at the light microscope level have shown that lactotrophs, somatotrophs, and thyrotrophs contain galanin in the intact female rat, whereas lactotrophs in the male AP do not. We recently reported that galanin and PRL release from estrogen-treated male and female pituitary cells in culture are coregulated by dopamine, TRH, and somatostatin. This suggested that galanin is stored within secretory granules, conceivably with PRL. Using postembedding immunocytochemistry at the ultrastructural level, the objectives of this study were to: 1) determine the subcellular location of galanin in the AP; 2) elucidate if galanin and PRL are colocalized within the same secretory granules; and 3) compare the cellular localization of galanin in the male and female AP. Male and ovariectomized female (OVEX) Fischer 344 rats were implanted with estradiol-containing or empty Silastic capsules for 2 weeks. Postembedding immunogold labeling was performed using rabbit (for galanin) and guinea pig (for PRL) generated antisera. Two different sizes of colloidal gold spheres were used to localize the hormones in the same tissue section. Galanin was primarily localized in secretory granules of adenohypophyseal cells. Based upon immunocytochemical results and morphological criteria, galanin was contained in somatotrophs but not lactotrophs in the male and OVEX AP. The AP of estrogen-treated rats contained more specific immunogold labeling for galanin than untreated rats. The increased immunoreactivity for galanin was notably associated with lactotrophs. After exposure to estrogen, galanin and PRL were colocalized within the same secretory granules of the male and OVEX pituitary cells. We conclude: 1) galanin is localized within secretory granules of the rat AP; 2) galanin and PRL are colocalized within secretory granules of the male and OVEX AP after estrogen treatment; and 3) galanin is localized in similar cell types in the male and OVEX AP, before and after estrogen treatment. These data provide a morphological basis for the coregulation of galanin and PRL secretion by hypothalamic factors.

Ultrastructural evidence for luteinizing hormone-releasing hormone neuronal control of estrogen responsive neurons in the preoptic area.

Both estrogen receptor (ER) immunoreactive (ir) and LHRH-ir neurons and processes are present in the preoptic area of the guinea pig. This experiment was conducted to determine if LHRH-ir terminals interact synaptically with ER-ir cells. A light microscopic dual chromogen immunocytochemical technique employing diaminobenzidine (DAB) and nickel-enhanced DAB for LHRH and ER localization, respectively, revealed that many varicose LHRH-ir fibers coursed in close proximity to ER-ir cells in the anterior part of the preoptic area at the preventricular periventricular nucleus (Pep), suggesting the likelihood of synaptic interactions. Ultrastructural analysis was performed using DAB and 3,3',5,5'-tetramethylbenzidine (TMB) for LHRH and ER localization, respectively. DAB labeling in LHRH-ir neurons appeared as a dense flocculent product dispersed throughout the cytoplasm. TMB stained ER-ir neurons contained electron dense crystalline spicules located predominantly in their nuclei. Numerous TMB labeled ER-ir neurons were present in the Pep, and occasionally occurred in clusters, closely apposed to one another. Many LHRH-ir terminals made synaptic contact or were apposed to unlabeled dendrites, while fewer contacted perikarya. Most significantly, ER-ir neurons showing clear evidence of intranuclear TMB crystals received synaptic input from LHRH-ir terminals. In addition, LHRH-ir terminals lacking synaptic specializations were also in direct apposition to ER-ir perikarya. These results provide morphological evidence that LHRH-ir neurons can regulate ER-ir neurons in the preoptic area. Since LHRH-ir cells are thought to be regulated by estrogen responsive neurons, interaction of LHRH terminals with ER-ir cells may represent a regulatory feedback circuit between the two systems.

Ultrastructural evidence for serotonin-immunoreactive terminals contacting phrenic motoneurons in the cat.

The innervation of the phrenic motor nucleus in the cat by serotonin-containing neurons has been studied using retrograde tracing combined with immunohistochemistry at the electron microscope level. It was found that phrenic motoneuron cell bodies and dendrites are contacted by serotonin-immunoreactive synaptic terminals. This finding suggests that the activity of phrenic motoneurons is directly affected by serotonergic neurons.

Ultrastructural identification of neuropeptides in the central nervous system.

A number of different neuropeptides have been described within presynaptic terminals at the ultrastructural level in the central nervous system. The majority of these neuropeptides share a common morphology with one another. Terminals containing neuropeptides have a population of small, clear vesicles associated with the active zone of the synapse and a lesser number of large, granular vesicles that are located at a distance from the active site of the synapse. It is believed that the large, granular vesicles act as a mechanism for the transport/storage of the neuropeptides, while the small, clear vesicles are thought to be acting as structures responsible for the release of the neurotransmitter/neuropeptide into the synaptic cleft. The neuropeptide containing terminals most often have asymmetrical junctions associated with their presynaptic membranes, although symmetrical junctions have been described with peptide containing terminals in a number of areas in the central nervous system. Neuropeptide containing terminals contact every part of the neuronal membrane; however, the majority of synaptic contacts involve portions of the dendritic shafts. Evidence is beginning to accumulate to indicate that for certain neuropeptides there is a specific spatial arrangement to their termination along the neuronal membrane.

Monoamine synaptic structure and localization in the central nervous system.

The monoamines dopamine, noradrenaline, adrenaline, and serotonin as well as the diamine histamine have a widespread distribution in the central nervous system within synaptic terminals and nonsynaptic varicosities. In certain regions of the central nervous system the monoamines are contained in varicosities that have no synaptic specialization associated with them, suggesting a possible neuromodulatory role for some of the monoamines. The majority of monoamine labelled structures are synaptic terminals which are characterized by the presence of small, clear vesicles (40-60 nm) and large, granular vesicles (70-120 nm) within the terminal. A third population of vesicles--small, granular vesicles--which are visible only after histochemical staining, are probably the equivalent of the small, clear vesicles present after either autoradiographic or immunohistochemical labelling. Most monoamine containing terminals contact dendrites and dendritic spines and, less frequently, neuronal somata and other axons. Both asymmetrical and symmetrical membrane specializations are associated with monoaminergic terminals; however, asymmetrical contacts are the most frequent type found. These ultrastructural results indicate that monoamine containing terminals and varicosities in general share many common morphological features, but still have diverse functions.

Diminished splenic function in asphyxiated term infants.

The effect of birth asphyxia on splenic function has not previously been investigated. We performed quantification of pocked erythrocytes to assess splenic function in 10 term infants with abnormally low Apgar scores and clinical symptoms of birth asphyxia and in 10 healthy term control infants. The mean pocked erythrocyte count of the asphyxiated infants at day 1 (33.9% +/- 6.8 [SD]) was greater than that of the control infants (21.4% +/- 6.3) (P less than .01). By day 3 the mean pocked erythrocyte count of the asphyxiated infants (25.2% +/- 8.4) was still elevated but approaching that of the control infants (18.8% +/- 5.5) (.05 less than P less than .10). In summary, birth asphyxia was associated with an elevation of pocked erythrocytes and decreased splenic function.

Polyamines and epidermal growth factor in monocrotaline-induced pulmonary hypertension.

Multiple lines of evidence suggest that the polyamines, a family of low-molecular-weight organic cations with documented regulatory note in cell growth and differentiation, are involved with hyperplastic and hypertrophic responses of lung cells underlying hypertensive pulmonary vascular disease. Little is known, however, of the factor(s) initiating polyamine synthesis in pulmonary hypertension. This study tested the key aspects of the hypothesis that augmented polyamine synthesis, and attendent vascular structural alterations in monocrotaline (MCT)-treated rats can be ascribed to elaboration of an epidermal growth factor (EGF)-like mitogen. In lungs of rats treated 4 days previously with 60 mg/kg, EGF-like immunoreactivity was detected diffusely throughout perivascular regions. Intravenous administration of human recombinant EGF (125 pg/h) to rats for 1 wk was associated with medial thickening in pulmonary arteries between 100 and 200 microns in diameter, significant increases in lung polyamine contents, and a moderate elevation in mean pulmonary arterial pressure. These observations indicate that EGF can be detected in the lungs of MCT-treated rats and that exogenous EGF mimics some of the action of MCT on the rat lung. It is thus reasonable to speculate that an EGF-like mitogen may participate in the response to MCT in part through a polyamine-dependent mechanism.

Shapes and projections of neurons with immunoreactivity for gamma-aminobutyric acid in the guinea-pig small intestine.

The distribution of nerve cell bodies and fibres with immunoreactivity for gamma-aminobutyric acid (GABA) has been studied in the guinea-pig small intestine. Cell bodies were common in myenteric ganglia but were extremely rare in the submucosa. Reactive fibres were numerous in the tertiary component of the myenteric plexus and in the circular muscle but they were rare in both myenteric and submucous ganglia. Reactive nerve fibres were absent from the mucosa. This distribution conforms to previous descriptions. Exposure to exogenous GABA, in vitro, was used to supplement endogenous stores of GABA. The morphology of cell bodies was better defined after this treatment. Nearly all cell bodies had type-I morphology, i.e., the cells had numerous short lamellar dendrites and one axon. Most axons ran anally. Some could be traced to the tertiary component of the myenteric plexus, others to the circular muscle. Removal of the myenteric plexus from a short length of intestine caused a loss of nerve fibres from the circular muscle beneath the site of operation and a decrease in fibre density in the circular muscle that extended anally from the lesion for about 1 mm. The nerve lesions caused no significant changes in the tertiary plexus. It is concluded that GABA is contained in motor neurons supplying the longitudinal and circular muscle, and that the neurons supplying the circular muscle may be inhibitory.

Cholecystokinin octapeptide immunoreactivity in the nucleus tractus solitarius of the cat.

The nucleus tractus solitarius possessed distinct patterns of cholecystokinin immunoreactive fibers and cell bodies within its various subdivisions. The commissural, medial, intermediate, parvocellular, dorsolateral and interstitial subdivisions contained relatively dense amounts of CCK immunolabelled fibers. In contrast, CCK immunoreactivity within the ventrolateral subdivision consisted of a few scattered fibers and small neurons. The commissural, intermediate, medial, dorsolateral and parvocellular subdivisions contained CCK immunoreactive neurons following colchicine treatment. The presence of CCK in the NTS suggest that it may be involved as a neuromodulator and/or neurotransmitter in circuitry that mediate cardiovascular, respiratory, gastrointestinal and taste functions.

Ornithine decarboxylase activity and immunohistochemical location in postischemic brain.

Ornithine decarboxylase, rate-limiting in polyamine formation, has been found to be necessary for the development of vasogenic edema after cryogenic cerebral injury and is postulated to be of importance in late ischemic brain edema formation. Ornithine decarboxylase activity and accompanying edema was studied after transient cerebral ischemia in Mongolian gerbils. Bilateral carotid artery occlusion was utilized to produce dense forebrain ischemia. After 4 h of reperfusion a significant elevation in ornithine decarboxylase activity was present (72.5 +/- 24.7 vs 8.5 +/- 2 pmoles/mg protein/h, p less than 0.05). Immunohistochemical localization of ornithine decarboxylase indicated its presence in cortical neurons of ischemic gerbils. This was typically located in the perinuclear cytoplasm and extended into proximal dendrites. Nonischemic animals did not contain ornithine decarboxylase immunoreactivity. These studies show the presence and location of ornithine decarboxylase in cerebral tissue subjected to transient ischemia. The increase in this marker of polyamine activity paralleled previous studies in this model of cerebral edema formation and reperfusion deficit in blood flow and evoked potential, suggesting that ornithine decarboxylase is a marker for and may be associated with those late metabolic events leading to progressive functional deterioration after incomplete cerebral ischemia.

Substance P binding sites in the nucleus tractus solitarius of the cat.

Substance P binding sites in the nucleus tractus solitarius were visualized with receptor autoradiography using Bolton-Hunter [125I]substance P. Substance P binding sites were found to have distinct patterns within the cat nucleus tractus solitarius. The majority of substance P binding sites were present in the medial, intermediate and the peripheral rim of the parvocellular subdivisions. Lower amounts of substance P binding sites were present in the commissural, ventrolateral, interstitial and dorsolateral subdivisions. No substance P binding sites were present in the central region of the parvocellular subdivision or the solitary tract. The localization of substance P binding sites in the nucleus tractus solitarius is very similar to the patterns of substance P immunoreactive fibers previously described for this region. Results of this study add further support for a functional role of substance P in synaptic circuits of the nucleus tractus solitarius.

Enkephalin-immunoreactive neurons in the nucleus tractus solitarius project to the parabrachial nucleus of the cat.

Enkephalin immunoreactive neurons within the nucleus tractus solitarius (NTS) were found to project to the parabrachial nucleus of the cat with the use of a combination of immunohistochemistry and retrograde transport of horseradish peroxidase. Double labelled neurons were located in the medial, parvocellular and commissural subdivisions of the NTS and were present predominantly ipsilateral to the injection site within the parabrachial nucleus. Only a few double labelled neurons were found in the contralateral NTS. The presence of neurons containing enkephalin immunoreactivity suggests that the role of enkephalin in the regulation of autonomic functions may be, in part, by circuits between the NTS and the parabrachial nucleus.

Immunohistochemistry of choline acetyltransferase in the guinea pig brain.

Choline acetyltransferase (ChAT) was localized immunohistochemically within the brain of the guinea pig using a monoclonal antibody. ChAT was found in the cytoplasm of cell bodies and primary dendrites of neurons located in striatum, basal forebrain, cranial nerve motor nuclei and scattered cells in the pons. The greatest numbers of immunoreactive neurons were located in the diagonal band of Broca, medial septum and striatum. Distinct immunoreactive fibers were not visible using this antibody, although a diffuse immunostaining was present in the same nuclear regions as well as in the nerve roots of cranial nerve nuclei and the interpeduncular nuclei. Results of the present study agree closely with other previous reports of acetylcholine distributions.

Colocalization of prolactin and growth hormone within specific adenohypophyseal cells in male, female, and lactating female rats.

The localization of PRL and GH within adenohypophysial cells has been investigated with immunocytochemical methodology using colloidal gold of different sizes. Classically, using morphological criteria at the light and electron microscope levels, two types of individual cells have been described which, it is believed, exclusively produce either PRL or GH, i.e. mammotrophs and somatotrophs. Since it has been reported that some gonadotrophs may colocalize and secrete both FSH and LH, and that some unidentified cells release both PRL and GH, we designed a study to investigate whether PRL and GH are present only in their respective specific cells or may be colocalized in mammotrophs and somatotrophs or possibly other types of cells within the pars distalis. Using immunocytochemistry at the electron microscope level we were able to label the two hormones with different sizes of colloidal gold bound to a second antibody and visualize them within pituitary cells. Two different primary antibodies to PRL and GH as well as a more purified antibody to GH were used. Pituitaries from cycling and lactating female rats and adult males were processed appropriately for electron microscopic studies. After sectioning, individual grids were treated with either primary antibodies for PRL or GH, or the same grid was treated sequentially with these two antibodies. All primary antibodies were absorbed with the heterologous hormone before usage. The second antibodies were bound to colloidal gold particles of either 10 nm (for GH) or 20 nm (for PRL) diameter, so that selective visualization of the two hormones could be achieved within the same cells. It was observed that mammotrophs immunolabeled only for PRL, and somatotrophs labeled only for GH regardless of the source of the antibody. However, an atypical small granule cell, a possible mammosomatotroph, colabeled consistently for both PRL and GH in all types of animals used and with antibodies from all of above sources. This report gives for the first time morphological evidence for the existence of pituitary cells that colocalize both PRL and GH in the normal rat pituitary gland. The possibility that these bihormonal cells represent stem cells which may give rise to both mammotrophs and somatotrophs under appropriate stimulation or that they may release both hormones under the influence of unspecific stimuli is suggested.