Real-time monitoring of estrogen-regulated gene expression in single, living breast cancer cells: a new paradigm for the study of molecular dynamics.

Elucidation of the molecular mechanisms underlying the development and progression of breast cancer has been hindered by two considerations. First, mammary tumor cells exhibit a considerable degree of morphological and functional heterogeneity. Thus, the conventional strategy of measuring a population response may not provide an accurate reflection of the behavior of the functional unit: an individual cell. Second, important regulatory events are generally separated in time, yet the strategies we use to monitor them are usually static as opposed to dynamic. With these considerations in mind, it would appear that a system for studying this problem should ideally combine the resolving power of single-cell analysis with a dynamic paradigm for making multiple measurements of gene expression from the same, living cell. This report summarizes our efforts at developing, validating, and optimizing such a system for monitoring hormone-driven gene expression in T47-D human breast cancer cells.

Mammosomatotropes: current status and possible functions.

Recent studies show that mammosomatotropes-cells that secrete both growth hormone and prolactin-are common in normal pituitary tissue. It is proposed that these dual hormone secretors serve as transitional cells for the differentiation of classical mammotropes during development, and for the functional interconversion of growth hormone and prolactin secretors in adults.

Dynamics of stimulus-expression coupling as revealed by monitoring of prolactin promoter-driven reporter activity in individual, living mammotropes.

Single-cell paradigms have greatly expanded our knowledge about stimulus-secretion coupling, but the understanding of stimulus-gene expression coupling has lagged behind for lack of a dynamic model sufficiently sensitive to provide single-cell resolution. In the present study, we made continuous indirect measurements within individual, living cells of expression dynamics both before and after treatment with a gene-activating secretagogue. This was accomplished by transfecting (via microinjection) individual, primary mammotropes with a PRL promoter-driven luciferase reporter plasmid, and then quantifying the rate of photonic emissions (reflective of endogenous gene activity). We found that individual cells exhibit spontaneous, random, short-term fluctuations of basal reporter activity and are extremely heterogeneous in terms of responses to a stimulatory agent (TRH). In addition, we found that responses are affected by several factors including the secretory status of the pituitary donor, the manner in which the stimulus is presented, and by the initial level of reporter activity. Moreover, the responsiveness of an individual cell can fluctuate dramatically over time. These results invite speculation that a given cell can "sense" its gene activation state and regulate its response accordingly to satisfy requirements for the corresponding secretory product.

Dynamic changes in spontaneous intracellular free calcium oscillations and their relationship to prolactin gene expression in single, primary mammotropes.

Cytosolic calcium plays a critical role in the control of a number of genes, including that of the pituitary hormone PRL. Cells that secrete this hormone, termed mammotropes, display spontaneous oscillations of intracellular free calcium ([Ca2+]i) that are positively correlated to PRL release. However, the precise contribution of calcium signaling to the expression of any gene including PRL has remained obscure owing to the requirement for and lack of a strategy for monitoring both of these dynamic variables (gene expression and [Ca2+]i oscillations) in the same living cell. In the present study, we overcame this technical limitation by making real-time measurements of PRL gene expression in transfected, primary rat mammotropes previously subjected to [Ca2+]i determinations by digital imaging fluorescence microscopy of fura-2. Our results showed that the majority of mammotropes (75%) exhibited distinct oscillatory behaviors that could be subgrouped on the basis of frequency/amplitude of [Ca2+]i changes, whereas the remainder (25%) were quiescent (nonoscillatory). Interestingly, most mammotropes displayed spontaneous transitions between oscillatory and quiescent states over the course of several hours. As a consequence of this oscillatory plasticity, there was not a positive correlation between [Ca2+]i dynamics and gene expression at any point in time, as would be predicted by studies with entire populations of cells. Instead, the relationship was distinctly inverse, suggesting that dynamic changes in PRL gene expression may be regulated by temporally dissociated transitions between quiescent and oscillatory states.

Dynamic monitoring and quantification of gene expression in single, living cells: a molecular basis for secretory cell heterogeneity.

Progress in understanding the dynamics of gene expression has been hampered by lack of a strategy for continuously monitoring this process within normal, living cells. Here, we employed a modification of conventional luciferase technology to make single and repeated real-time measurements of PRL gene expression from individual, living lactotropes from nursing rats. Cells were individually transfected by microinjection with a PRL promoter/luciferase reporter construct. Levels of PRL gene transcription were quantified by photonic imaging in the same cells before and after 24 h of culture in the presence or absence of the dopamine agonist bromocryptine or epidermal growth factor, two well known regulators of PRL gene transcription. We found these cells to be remarkably heterogeneous with respect to basal PRL gene expression and that the degree of activity within a single cell could fluctuate greatly over time under basal culture conditions. Treatment with bromocryptine or epidermal growth factor induced predictable and reversible changes in the average responses observed, yet individual cells displayed marked differences in response to these agents. These findings demonstrate the utility of this paradigm for monitoring dynamics of gene expression within normal, living cells of any type. Moreover, they provide a molecular basis for the secretory heterogeneity and plasticity that have come to be known as hallmarks of lactotrope cell function.

Spontaneous calcium oscillatory patterns in mammotropes display non-random dynamics.

We previously showed that primary rat mammotropes exhibited four distinct patterns of 'spontaneous' free intracellular calcium ([Ca2+]i) oscillatory behavior: a quiescent state A and three oscillatory states B,C&D, which differed in frequency/amplitude characteristics. When [Ca2+]i was monitored in 10 min windows separated by several hours, these phenotypes were frequently found to interconvert, raising the question about whether these transitions were random or ordered events. We reasoned that if such activity were random, then neither episode duration nor transitional probabilities should differ among phenotypes. We tested this logic in the current study by making long-term, continuous measurements of [Ca2+]i in mammotropes microinjected with Fura-2-dextran and identified by their ability to express a prolactin promoter-driven reporter plasmid. We found that transitions occurred in ~25% of cells (n = 36 from 9 independent experiments) once every 1-5 h and demarcated phenotype episodes of different duration (A, 1.04 +/- 0.2 h; B, 1.64 +/- 0.3 h; C, 2.45 +/- 0.62 h; D, 0.90 +/- 0.2 h, mean +/- SEM). Moreover, some transitions occurred more frequently than others and linked specific phenotypes into a common pattern: C to B to A. Our results demonstrate that the seemingly spontaneous nature of [Ca2+]i phenotype transitions are, in fact, ordered and support the view that they comprise a structured 'code' like that proposed to underlie calcium-dependent regulation of exocytosis and gene expression.

Ovine prolactin (PRL) and dopamine preferentially inhibit PRL release from the same subpopulation of rat mammotropes.

Autoregulation of PRL release was studied at the single cell level by the use of a reverse hemolytic plaque assay. Monodispersed pituitary cells from adult male rats were first preincubated with test substances and then coincubated with antirat PRL antiserum before development of plaques with complement. At the conclusion of the assay, the percentage of all pituitary cells in culture that formed plaques was evaluated microscopically, and the rate of plaque development was used as an index for the rate of hormone release. In controls, the maximal percentage of pituitary cells formed PRL plaques within a 1.5-h antibody incubation period, and addition of TRH (1 X 10(-7) M) did not increase this proportion. Treatment with ovine PRL (oPRL, 100 ng/ml) or dopamine (1 X 10(-7) M), either alone or in combination, caused a comparable suppression of the rate of PRL plaque development, which was reversed by the presence of TRH. Pretreatment of cells with the lysosomotropic agent chloroquine (1 X 10(-5) M) overrode dopamine inhibition of PRL plaque development, but did not influence oPRL inhibition. Taken together, these results demonstrate that oPRL inhibits basal, but not TRH-induced, PRL release from rat pituitary cells and support the view that PRL can act at the pituitary level to inhibit its own secretion. Moreover, the equipotency and lack of additivity exhibited by oPRL and dopamine coupled with the differential effects of chloroquine suggest that these factors both act upon the same subpopulation of mammotropes to inhibit PRL release, but by separate intracellular mechanisms.

Hypothalamic factors differentially affect the proportions of cells that secrete growth hormone or prolactin.

Hypothalmic factors have been shown to regulate acutely the synthesis and release of adenohypophyseal hormones, yet few studies have investigated the long term effects of these agents on adenohypophyseal cell types. In the present study, we assessed the chronic influence of selected hypothalamic factors on the relative proportions of GH- and PRL-secreting cells in pituitary cultures derived from 5-day-old rats. Primary cultures were established and incubated for 6 days in the presence or absence of 0.1-microM doses of GH-releasing factor, LHRH, CRF, TRH, or vasoactive intestinal polypeptide and then subjected to reverse hemolytic plaque assays for analysis of the percentages of cells that released GH or PRL. In cultures from males, GH-releasing factor and LHRH treatment caused an increase in the proportion of PRL secretors and a commensurate decrease in the GH population. CRF increased PRL cells without affecting the GH secretors, while TRH reduced the percentages of both cell types, and vasoactive intestinal polypeptide had no effect. Virtually identical results were obtained for cells isolated from females. These results demonstrate that hypothalamic factors have the capacity to induce differential effects on the proportions of GH- and PRL-secreting cells. Surprisingly, our findings also show that hypothalamic factors that do not normally influence the acute release of GH or PRL can exert a chronic effect on the proportion of cells that secrete these hormones.

Functional variations among prolactin cells from different pituitary regions.

Reverse hemolytic plaque assays were used to compare the responsiveness of cells from different pituitary regions to the modulatory effects of human pancreatic GH-releasing factor (GRF), TRH, and dopamine (DA). Tissues from the peripheral rim (outer zone) and the central region (inner zone) of adenohypophyses from day 10 lactating rats were dispersed with trypsin, and the cells were placed into culture. On the following day, these cells were subjected to GH plaque assays (conducted in the presence or absence of GRF) and PRL plaque assays (performed with or without TRH and DA). Cells from both zones responded similarly to GRF with a rapid acceleration of GH plaque formation. However, the rate of PRL plaque formation in response to TRH and DA differed between cells from these regions. For outer zone cells, plaque development increased greatly with TRH treatment, but was only moderately affected by DA. Plaque formation from inner zone cells was influenced slightly by TRH, but markedly inhibited by DA. These results suggest that PRL, but not GH, cells from these pituitary regions are differentially responsive to at least two hypothalamic secretagogues. We then performed fixed sequential plaque assays to determine whether the proportions of cells that released PRL only (classical mammotropes) or those that released both GH and PRL (mammosomatotropes) also differed between the inner and outer zones. Using this approach, we found that the outer zone contained a much larger proportion of dual hormone secretors than did the inner zone. These results, when taken together with the responsiveness differences discussed above, raise the possibility that the release of PRL from mammotropes and mammosomatotropes is regulated differently and that the ratio of these two cell types may dictate, in part, the manner in which a specific region of the pituitary responds to hypothalamic input.

Neurointermediate lobe peptides recruit prolactin-secreting cells exclusively within the central region of the adenohypophysis.

There is strong evidence that the hypophyseal neurointermediate lobe (NIL) mediates 17 beta-estradiol (E2)-induced PRL secretion in rats. Our laboratory has previously demonstrated that E2 stimulates NIL cells to release an activity that acutely increases the relative abundance of PRL-releasing cells in anterior pituitary (AP) cell cultures. We later found that secretory products of the NIL melanotropes, specifically the acetylated forms of alpha MSH and beta-endorphin (beta END), can account for this activity. Given that blood from the NIL initially perfuses the region of the AP proximal to the NIL, we tested the hypothesis that this specific area was preferentially responsive to the lactotrope recruitment activities. AP glands from ovariectomized rats were dissected into an inner zone, proximal to the NIL, and the remaining outer zone of the gland, then dispersed with trypsin. The resulting cells were cultured for 16 h, either alone or in coculture with NIL cells, and subsequently treated with medium alone (control) or with alpha MSH, beta END, or E2 (all at 1 x 10(-7) M) for 3 h and then subjected to reverse hemolytic plaque assays for PRL release. Under control conditions, the proportion of PRL-secreting cells was significantly greater in cultures from the outer zone of the AP than in those from the corresponding inner zone of the gland. Treatment of AP cells from the inner zone with alpha MSH, beta END, or the E2-induced NIL activity significantly increased the percentage of PRL secretors by about 8% of all AP cells. In contrast, the fraction of PRL-secreting cells in cultures from the outer zone was not affected by these treatments. We conclude that the recruitment of PRL-secreting cells in response to products of the NIL occurs only in that region of the AP proximal to the NIL.

Biphasic effects of estrogen on gonadotropin-releasing hormone-induced luteinizing hormone release in monolayer cultures of rat and monkey pituitary cells.

The negative feedback effect of estrogen on LH secretion has been difficult to demonstrate in monolayer cultures of rat pituitary cells. The purpose of the present study was to establish the experimental conditions required for manifestation of this response and, in the process, to develop models for investigating the actions of steroids on pituitary cells. Both dynamic and static incubation systems were used. For perifusion experiments, trypsin-dispersed pituitary cells from rats at random stages of the estrous cycle were attached to glass coverslips with poly-L-lysine, incubated for 48 h, and then mounted in Sykes-Moore chambers. In each of these experiments, two chambers were perifused concurrently: one with medium containing 1.8 X 10(-10) M 17 beta-estradiol and the other with medium alone. GnRH (4.2 X 10(-9) M) was coinfused for 5 min out of every hour, and samples of perifusate were collected as 5-min fractions for assay of LH. Estrogen treatment significantly (P less than 0.01) suppressed LH release in response to the first five GnRH pulses compared to the control value. The inhibition was most pronounced early in the perifusion, but had disappeared by 6 h. These results demonstrate that estradiol exerts a potent but transient inhibition of GnRH-induced LH release in monolayer cultures of rat pituitary cells. In a subsequent set of experiments, we modified a static incubation system to assess sequentially the biphasic effects of estrogen on LH release by the same group of cells. Cultures of rat pituitary cells that had been established 42 h previously were treated simultaneously for 3 h with 17 beta-estradiol (3.7 X 10(-10) M) and various concentrations of GnRH (5 X 10(-10) to 1 X 10(-7) M) to measure the inhibitory effects of the steroid on LH secretion. This experiment was repeated on the same cells after 27 h of steroid exposure to estimate the facilitory actions of estrogen on LH release. The negative feedback of estrogen was demonstrable in static cultures of rat pituitaries provided that the period of estrogen exposure and duration of incubation were brief. Moreover, the results indicate that the same groups of cells can be used on consecutive days to investigate the inhibitory and stimulatory effects of estrogen on LH secretion. Experiments with cultures of monkey pituitary cells yielded similar results. Taken together, these findings indicate that cultured pituitary cells are responsive to the biphasic actions of estradiol and demonstrate the utility of two model systems for investigating these phenomena.

Liver tissue produces a potent lactogen that partially mimics the actions of prolactin.

An in vitro bioassay (based on the detection of casein release from isolated mammary cells by reverse hemolytic plaque assay) was used to detect a lactogenic factor secreted by liver tissue from suckled rats. This material stimulates casein release in the absence of added PRL and is actually more potent than PRL in this regard. The substance(s) possesses the following additional characteristics: it is released from liver tissue of lactating but not virgin female or male rats; when tested together with PRL, its effects are additive rather than synergistic; and unlike PRL, it does not increase the proportion of mammary cells committed to casein release. These findings are consistent with the possibility that a Liver Lactogenic Factor may mediate, at least in part, the lactogenic component of PRL's action. This functional relationship may be similar to that which exists between GH and the somatomedins.

Capacity of individual somatotropes to release growth hormone varies according to sex: analysis by reverse hemolytic plaque assay.

A reverse hemolytic plaque assay was used to investigate the mechanism(s) underlying sexual differences in GH release which evolve at puberty in rats. The percentages of GH-secreting cells in 24-h pituitary cultures from each sex were similar for pituitary donors up to 30 days of age (range = 38.9% to 41.7% of all cells in culture, n = 3 separate experiments) but decreased by day 50. The decrease was more striking for females (to 24.1 +/- 0.3% mean +/- SE) than for males (to 33.2 +/- 1.1%). However, owing to the greater increase in total pituitary cell number exhibited by female rats at this time, the absolute numbers of somatotropes recovered from male and female pituitaries were almost identical on 50 and 100 days of age. To assess the secretory capacities of individual somatotropes, we measured the sizes of plaques formed. In prepubertal rats (days 10-30), the plaque areas under basal conditions were comparable for males and females at each age studied, and treatment with GH-releasing factor increased plaque sizes to approximately the same degree (10-fold) for both sexes at each age. However, by day 100, plaques that formed under both basal and stimulated conditions were consistently larger (P less than 0.01) for male than for female donors. Taken together, our results demonstrate that sexual differences in GH release are attributable to the secretory capacities of individual somatotropes rather than to differences in the numbers of GH cells in pituitaries of male and female rats.

Ontogeny of prolactin secretion in the neonatal rat is regulated posttranscriptionally.

PRL-secreting cells first appear in appreciable numbers on day 4 of neonatal life in rats. In the present study, we attempted to ascertain whether the ontogenic appearance of PRL mRNA and hormone release were temporally coordinated or discordant. Our results show that the PRL gene is expressed at least 3 days before the onset of secretion in newborn rats. Moreover, steady state levels of PRL mRNA in neonates are at least as high as those found in 10-day-old rats, in which 15-17% of all pituitary cells secret the hormone. This apparent blockage of translation is attributable to a lack of association of PRL message with ribosomes in the neonate. Taken together, these data indicate that the ontogeny of PRL secretion in rats is regulated translationally as well as transcriptionally.

Suckling increases the proportions of mammotropes responsive to various prolactin-releasing stimuli.

It is well established that the suckling stimulus sensitizes or primes the anterior pituitary to PRL-releasing stimuli. It is also recognized that PRL-secreting cells from a given animal are not all alike but instead exhibit a considerable degree of functional heterogeneity. The goal of the present study was to determine whether the suckling-induced priming phenomenon is manifest at the cellular level by shifts in the relative abundance of various mammotrope subpopulations. This was accomplished by using reverse hemolytic plaque assays to evaluate the secretory characteristics of individual PRL secretors derived from lactating rats either before or after the transient application of a suckling stimulus. Groups of day 10 lactating rats separated from their litters for 4 h were either killed immediately or were reunited briefly (10 min) with their pups before death. Adenohypophyseal cells obtained after trypsin dispersion were then subjected to plaque assays for PRL. Mammotropes derived from suckled rats were, on average, considerably more responsive to the stimulatory actions of TRH and angiotensin II and less susceptible to inhibition by dopamine. Mammotropes from nonsuckled rats exhibited a bimodal frequency distribution in which plaques from the second mode were roughly 6-8 times larger (released considerably more PRL) than those from the first. Superimposition of suckling (or in vitro treatment with dopamine) caused the second mode to disappear. Suckling also enhanced greatly the fraction of PRL cells that shifted from the first to the second mode (i.e. released more hormone) after treatment with TRH or angiotensin II. Taken together, our results demonstrate that the suckling-induced sensitization of pituitary tissue to PRL-releasing stimuli is manifest at the cellular level as proportional shifts toward those cells most responsive to stimulatory secretagogues and away from those most susceptible to inhibition by dopamine.

Stimulation of prolactin cell differentiation in vitro by a milk-borne peptide.

We have previously reported that the normal expression of PRL-secreting cells in neonatal rats requires a maternal signal specific to the first few days of lactation. These results raised the possibility that a milk-borne factor(s) ingested by the neonate and absorbed into the circulation might induce the ontogenic appearance of PRL cells. The purpose of the present study was to determine whether milk from this period could directly stimulate the differentiation of PRL secretors in culture. Monodispersed anterior pituitary cells from 1-day-old pups were cultured for 6 days with aqueous extracts of milk from early (days 2, 3, and 4) and late (days 15 and 16) lactation and then subjected to reverse hemolytic plaque assays for PRL and GH release. We found that the addition of milk extracts (10 mg/ml) from either early or late lactation stimulated the differentiation of PRL secretors (to 6.1 +/- 1.0% and 2.4 +/- 0.7% of all pituitary cells, respectively; mean +/- SE; n = 3) above that in control cultures without milk (0.2 +/- 0.2%). Thus, early milk was more than twice as effective as late milk in this regard (P less than 0.05). This effect appeared to be specific to PRL cell differentiation, since the relative abundance of GH secretors was not different between cells treated with either early or late milk (29.3 +/- 4.8% and 33.7 +/- 3.9%, respectively). On the other hand, late milk was more than twice as effective as early milk at increasing the capacity of GH secretors to release hormone (P less than 0.05). Preliminary characterization by gel filtration chromatography and proteolytic hydrolysis indicates that the bioactivity that differentiates PRL secretors is a small peptide(s) of 2000-6000 daltons. Taken together, our results demonstrate that a milk-borne peptide(s) is capable of specifically stimulating the differentiation of PRL-secreting cells in vitro, and that this bioactivity is more prevalent in milk from early lactation.

Detection of hormone release from individual cells in mixed populations using a reverse hemolytic plaque assay.

Prolactin (Prl) secreting cells in a mixed pituitary cell culture form microscopically-identifiable plaques (zones of hemolysis around the lactotropes) when incubated in a monolayer with staphylococcal protein-A-coated ovine erythrocytes in the presence of Prl antiserum and complement. Plaques form first at 15-30 min and are maximal in size and number at 2 h. Approximately 70% of the adenohypophyseal cells form plaques under these conditions. TRH increases, and dopamine decreases, the size and number of plaques at early times during incubation. This reverse hemolytic plaque assay probably can be used to detect any cell secretion for which an antibody is available. This technique, or a modified version of it in which sequential plaque assays are performed on identified cells--used together with immunocytochemistry, autoradiography or electron microscopy of those cells--should provide better answers to commonly asked questions about secretory systems: Do all or only a subset of cells containing the same hormone respond to a particular secretagogue? Can cells that contain two hormones release one of them preferentially?

Mode of gonadotropin secretion in infantile female rats and the role of estrogen in feedback regulation.

The temporal aspects of gonadotropin release were studied in infantile (14-day-old) female rats. Blood samples were collected using vena cava cannulae at 10-, 15-, 20-, or 30-min intervals and assayed for either LH alone or for both LH and FSH. Within individual animals, blood LH levels exhibited a degree of variability suggestive of a pulsatile release of the hormone. Wide fluctuations in circulating FSH were also observed, indicating that FSH might also be secreted in a pulsatile manner. Peak values for both gonadotropins were 2- to 5-fold higher than baseline levels. The majority of pups exhibited a nonrhythmic release of hormones. In some animals, LH and FSH appeared to be secreted in a temporally coincident manner. Administration of LHRH to cannulated pups evoked a simultaneous discharge of both gonadotropins. To determine whether endogenous 17 beta-estradiol (E2) plays a physiological role in the control of LH secretion in infantile rats, circulating LH levels were monitored in cannulated pups which had been treated twice daily with anti-E2 serum for 4 days. Passive immunization to E2 did not affect pulsatile LH secretion. Therefore, the results of this study demonstrate that in infantile female rats, LH is secreted in a pulsatile manner which is independent of E2 feedback regulation.

Rapid augmentation of prolactin cell number and secretory capacity by an estrogen-induced factor released from the neurointermediate lobe.

17 beta-Estradiol (E2) has been shown to exert an acute stimulatory effect on PRL secretion via an indirect action involving the neurointermediate lobe (NIL). In the present study we used a reverse hemolytic plaque assay to determine whether this effect was manifested as an augmentation of the number of PRL secretors and/or an increase in the amount of hormone released per PRL cell. Cultures of anterior pituitary (AP) and NIL cells from ovariectomized rats were cultured overnight, exposed to the treatment (E2 or vehicle) for 3 h, and then subjected to a reverse hemolytic plaque assay that was carried out in the presence or absence of TRH. Concurrent exposure of AP cells to E2 and NIL cells evoked an 11-12% increase in the overall proportion of PRL-secreting cells. This was true when the AP and NIL cells were incubated as a mixed culture and when the two cell types were maintained in the same petri dish but on separate plastic supports, and TRH did not significantly influence this response. The effect of E2 on the number of PRL secretors was negated when NIL cells were not present throughout the experiment or when they were removed from the cultures just before commencement of E2 treatment. Simultaneous treatment with E2 and NIL cells also significantly augmented the sizes of PRL plaques produced under basal conditions by AP cells. Taken together, these results demonstrate that E2 stimulates NIL cells to release an activity that enhances PRL secretion in two ways: 1) by recruiting additional PRL cells into the secretory pool, and 2) by augmenting the secretory capacity of individual PRL cells.

Alpha-melanocyte-stimulating hormone is a mammotrophic factor released by neurointermediate lobe cells after estrogen treatment.

When neurointermediate lobe (NIL) cells from ovariectomized rats are exposed to 17 beta-estradiol (E2) in vitro, they release a substance that rapidly induces the recruitment of additional PRL cells into the secretory pool. In the present study we tested the hypothesis that this mammotrophic factor is alpha MSH. Cocultures of anterior pituitary and NIL cells were incubated for 18 h, exposed to various treatments for 3 h, and then subjected to a reverse hemolytic plaque assay to quantify the percentage of all pituitary cells that released PRL. Combined exposure to E2 and NIL cells caused a significant increase in the fraction of anterior pituitary cells that released PRL, and the presence of TRH during the reverse hemolytic plaque assay did not affect the magnitude of the response. Treatment with dopamine (which inhibits alpha MSH release) reversed the recruitment of PRL secretors induced by E2 stimulation of NIL cells. Likewise, immunoneutralization with an antiserum directed against alpha MSH abolished the response. Furthermore, alpha MSH alone could substitute for E2 and NIL cells in evoking the recruitment response, whereas none of several other POMC-derived peptides had any consistent effect. Taken together, these results demonstrate that alpha MSH is a mammotrophic factor released by NIL cells in response to E2.