Autocrine Positive Feedback Regulation of Prolactin Release From Tilapia Prolactin Cells and Its Modulation by Extracellular Osmolality

Abstract

Prolactin (PRL) is a vertebrate hormone with diverse actions in osmoregulation, metabolism, reproduction, and in growth and development. Osmoregulation is fundamental to maintaining the functional structure of the macromolecules that conduct the business of life. In teleost fish, PRL plays a critical role in osmoregulation in fresh water. Appropriately, PRL cells of the tilapia are directly osmosensitive, with PRL secretion increasing as extracellular osmolality falls. Using a model system that employs dispersed PRL cells from the euryhaline teleost fish, Oreochromis mossambicus, we investigated the autocrine regulation of PRL cell function. Unknown was whether these PRL cells might also be sensitive to autocrine feedback and whether possible autocrine regulation might interact with the well-established regulation by physiologically relevant changes in extracellular osmolality. In the cell-perfusion system, ovine PRL and two isoforms of tilapia PRL (tPRL), tPRL₁₇₇ and tPRL₁₈₈, stimulated the release of tPRLs from the dispersed PRL cells. These effects were significant within 5–10 minutes and lasted the entire course of exposure, ceasing within 5–10 minutes of removal of tested PRLs from the perifusion medium. The magnitude of response varied between tPRL₁₇₇ and tPRL₁₈₈ and was modulated by extracellular osmolality. On the other hand, the gene expression of tPRLs was mainly unchanged or suppressed by static incubations of PRL cells with added PRLs. By demonstrating the regulatory complexity driven by positive autocrine feedback and its interaction with osmotic stimuli, these findings expand upon the knowledge that pituitary PRL cells are regulated complexly through multiple factors and interactions.

Prolactin (PRL) is a vertebrate hormone produced by cells of the adenohypophysis and was originally named for its role in stimulating lactation in rabbits (1, 2). This hormone is reported to exert over 300 separate functions related to reproduction, osmoregulation, immunoregulation, growth, metabolism, brain function, development, and behavior (3). Appropriately, the regulation of PRL secretion is multifaceted and complex.

Osmoregulation is fundamental to maintaining the functional structure of the macromolecules that conduct the processes of life. For this reason, a substantial portion of the vertebrate endocrine system is dedicated directly or peripherally to regulating salt and water balance. In teleost fishes, PRL plays an indispensable role in maintaining osmotic homeostasis in fresh water (FW) due to its ability to promote ion retention and reduce water permeability (4–7).

PRL has become an important topic in oncogenic studies, as it stimulates cell proliferation (3, 8, 9). PRL has also been found to be produced in peripheral tissues, such as the breast and prostate gland (10). High levels of PRL are found in tumors within these tissues (11, 12) suggesting that PRL may induce tumor growth via paracrine/autocrine pathways. Meanwhile, the effects of PRL on the primary site of its production in the pituitary are not well understood.

In most vertebrates, the principal hindrance to studying the mechanisms that regulate PRL release is the great difficulty in obtaining homogeneous preparations of normal PRL cells, owing to their dispersal among other adenohypophyseal cell types. Furthermore, the high concentration of endogenous PRL in pituitary has been another obstacle for addressing the autocrine actions of PRL. In mammals, this problem has been partly solved through the employment of clonal PRL tumor cell lines (by definition, not normal) and by interfering with PRL receptor (PRLR) signaling (see Ref. 13). Ovine PRL (oPRL) was found to down-regulate the transcription and secretion of PRL in rat GH3, GC, and 2B8 clone cells (14–16), and to stimulate the internalization of PRLR in rabbit mammary gland, when injected (17). The enhanced internalization of PRLR was replicated in human embryonic kidney 293T cells and Jak2-null g2A cells with homologous PRL (18).

Recent studies employing a PRLR antagonist or PRLR-knockout mice suggest that, despite of its established cell-proliferative effects in many tissues, PRL acts as an antiproliferative factor in anterior pituitary cells, including PRL cells (19, 20). Such model systems, however, possess several limitations that include functional and morphological differences from normal PRL cells that restrict their general applicability (21–26). Nevertheless, insofar as these observations reflect normal mammalian PRL cells, they suggest that pituitary PRL cells are under negative autocrine regulation.

In the pituitary of the euryhaline teleost fish, the tilapia, Oreochromis mossambicus, well over 99% of the rostral pars distalis (RPD) is comprised by PRL cells (27). This nearly homogeneous population of tilapia PRL (tPRL) cells greatly facilitates in vitro studies on the regulation of PRL transcription, translation and secretion (see Refs. 28, 29). Consistent with its role in FW osmoregulation, the tPRL cell is a well-established osmoreceptor by virtue of its ability to alter the release of two PRL isoforms, which are encoded by separate genes, tPRL₁₇₇ and tPRL₁₈₈, in direct response to a change in extracellular osmolality (30–34). In Mozambique tilapia, the synthesis and release of both tPRL₁₇₇ and tPRL₁₈₈ rise and fall inversely with changes in extracellular osmolality (32); although both of these hormones possess similar ion-retaining activities (35), the gene expression and release of tPRL₁₈₈ respond more robustly in vivo and in vitro to hyposmotic stimuli than that of tPRL₁₇₇. Over the longer term, both tPRL cell size and the area occupied by tPRL cells in the RPD are greater in FW-acclimated fish than in seawater (SW)-acclimated fish (36). Due to their suitability for in vitro studies, osmoreceptive nature, and the ability to measure the gene expression of PRLRs (32), tPRL cells allow us to examine the autocrine regulation of PRL in normal cells along with possible interactions with variations in extracellular osmolality.

Here, we assessed the autocrine regulation of dispersed tPRL cells employing oPRL, which is often employed in fish models (37, 38), and both native tPRL isoforms. Taking into account the role played by PRL cells in osmoreception, we also tested the hypothesis that the potency and/or direction of actions of oPRL and the two tPRLs on the release and gene expression of tPRLs from dispersed tPRL cells is sensitive to changes in extracellular osmolality.

Materials and Methods

Animals

Mature Mozambique tilapia (O. mossambicus) of mixed sexes and sizes (31.7–968.8 g) were obtained from stocks maintained at the Hawai'i Institute of Marine Biology, University of Hawai'i. Fish were reared in outdoor tanks with a continuous flow of FW under natural photoperiod, and fed approximately 5% of their body weight per day with Silver Cup Trout Chow (Nelson and Sons, Inc). All experiments were conducted in accordance with the principles and procedures approved by the Institutional Animal Care and Use Committee, University of Hawai'i.

Dispersed tPRL cells and experimental media

Dispersed tPRL cells were prepared from the RPDs of O. mossambicus as previously described (32, 39) with minor modifications. Briefly, RPDs dissected out from FW-acclimated tilapia were pooled in PBS (0.02M, 355 mOsm/kg) and treated with 0.125% (wt/vol) trypsin (Sigma-Aldrich) in PBS for 25 minutes. After termination of trypsin treatment with trypsin inhibitor (0.1% wt/vol; Sigma-Aldrich), cells were resuspended in hyperosmotic medium (355 mOsm/kg; see below) and subjected to either static incubations or perifusion experiments. Cell viability and yield were determined using trypan blue and a hemocytometer. The number of RPDs used for static incubations with oPRL was 46, and that used for static incubations with tPRLs was 98 and 39 for low (0.01, 0.1, and 1 μg/mL) and high concentrations (10 μg/mL), respectively. For perifusion experiments, an average 4.61 RPDs were used per chamber.

The incubation media contained 120mM NaCl, 4mM KCl, 0.81mM MgSO₄, 0.99mM MgCl₂, 2mM NaHCO₃, 0.44mM KH₂PO₄, 1.34mM Na₂HPO₄, 2.1mM CaCl₂, 10mM HEPES, 2.77mM glucose, 2mM glutamine, 100-IU/mL penicillin, 76.3-IU/mL streptomycin, and DMEM. After the adjustment of media pH to 7.55, media osmolalities were checked using a vapor pressure osmometer (Wescor 5100C; Wescor) and adjusted to 355 mOsm/kg for hyperosmotic medium and 300 mOsm/kg for hyposmotic medium by adding 5M NaCl.

Source of hormones

oPRL was purchased from Sigma-Aldrich. tPRL₁₈₈ and tPRL₁₇₇ were purified from media after the incubation of RPDs of FW-reared tilapia (40, 41). Briefly, 1 mL of 1% acetic acid was added to 20–25 mL media before centrifugation to remove debris. Media was then passed through a Sep-Pak C18 cartridge (55–105 μm particle size; Nihon Waters), equilibrated with 0.1% trifluoroacetic acid (TFA). The absorbed proteins were then eluted with 80% acetonitrile in 0.1% TFA and lyophilized. Lyophilized samples were dissolved in 0.1% TFA and separated by reverse-phase HPLC (Gulliver; Jasco) on an ODS-120T column (0.46 × 25 cm, 5-μm particle size; Tosoh), employing 60 minutes linear gradient of 40%–60% acetonitrile in 0.1% TFA at 40°C. A flow rate was 1 mL/min and absorption was monitored at 220 nm. The HPLC-purified fractions were lyophilized, separated by SDS-PAGE, and verified by Western blotting. The proteins were visualized with 1% amido black in 7% acetic acid. The N-terminal amino acid sequences of proteins were determined using a protein sequence analyzer (Shimadzu PPSQ-10; Shimadzu). The sequences for the identification of tPRL₁₈₈ and tPRL₁₇₇ were VPINDLL and VPINDLI, respectively.

Static incubations

The dispersed tPRL cells were plated on 96-well plates (96 600, 156 000, and 155 000 cells per well for experiments with oPRL, and low [0.01, 0.1, and 1 μg/mL] and high concentrations [10 μg/mL] of tPRLs, respectively; 8–12 replicates per treatment) and incubated at 26°C under saturated humidity. After the preincubation for 1 hour in hyperosmotic medium, the cells were exposed to hyper-/hyposmotic media containing oPRL (0, 0.1, 1, and 10 μg/mL) or either tPRL₁₈₈ or tPRL₁₇₇ (0, 0.01, 0.1, and 1 μg/mL or 10 μg/mL). At the end of 6 hours incubation, media were collected and diluted 20 times with RIA buffer (0.01M PBS containing 1% [wt/vol] BSA and 0.1% [vol/vol[ Triton X-100), and stored at −20°C until analyzed. After media collection, 100 μL of TRI reagent (MRC) was added to each well. The cell samples with TRI reagent were then transferred to 1.5 mL tubes and stored at −80°C before RNA extraction and gene expression analyses.

Perifusion experiment

The dispersed tPRL cells were plated in poly-L-lysine (0.083 mg/mL; Sigma-Aldrich)-coated chambers, which were prepared as previously described (33). Up to 8 chambers, which varied in volume from 190 to 320 μL, were perifused in each experimental run using a peristaltic pump (Cole-Parmer 7618-60; Cole-Parmer) with flow rates of 250 ± 20 μL/5 minutes, via separate inlet tubings (Supplemental Figure 1). The average number of tPRL cells per chamber ranged from 480 000 to 1 569 000 among different runs; within a single run, variations in chamber volume, and consequently in number of tPRL cells per chamber, were limited to 5.19% of the average. Each treatment was tested with 4–8 replicates. The chambers were submerged in a water bath at 26.4°C to keep their internal temperature stable throughout the experiments. After 1 hour of preperifusion with hyperosmotic medium, the chambers were perifused for an additional 30 minutes to establish a baseline release of tPRLs, and then exposed to hyper- or hyposmotic media with or without oPRL (10 μg/mL), tPRL₁₈₈, or tPRL₁₇₇ (1 μg/mL). The flow-through was collected every 5min and stored at −20°C until analyzed.

Quantitative real-time PCR (qRT-PCR)

Total RNA was extracted from cells frozen in TRI reagent, following the manufacturer's protocol. The first-strand cDNA was synthesized from either 40 ng (experiment using 0.01–1 μg/mL of tPRLs) or 80 ng (experiments using oPRL and 10 μg/mL of tPRLs) of total RNA, using High Capacity cDNA Reverse Transcription kit (Thermo Fisher Scientific). The absolute quantity of transcripts of target and reference genes were determined by qRT-PCR using the StepOnePlus real-time PCR system (Thermo Fisher Scientific). The reaction mix (15 μL) contained Power SYBR Green PCR Master Mix (Thermo Fisher Scientific), 200nM forward and reverse primers, and cDNA. The standard curves for absolute quantification were generated using serially diluted standard cDNA samples, which were composed of target gene cDNA fragments of known concentration. The following cycling parameters were employed: 2 minutes at 50°C, 10 minutes at 95°C followed by 40 cycles at 95°C for 15 seconds and 60°C for 1 minute. All qRT-PCR primers used were previously reported and listed in Table 1 (42, 43). The measurements of target genes were normalized by those of elongation factor 1α as an internal reference and expressed as percent changes from control (355 mOsm/kg, without tested PRLs).

Table 1.

List of qRT-PCR Primers

Target Gene	Forward Primer (5′-3′)	Reverse Primer (5′-3′)	Reference
EF1α	AGCAAGTACTACGTGACCATCATTG	AGTCAGCCTGGGAGGTACCA	42
tPRL188	GGCCACTCCCCATGTTTAAA	GGCATAATCCCAGGAGGAGAC	43
tPRL177	TGGTTTGGCTCTTTTAACACAGTG	AGACAATGAGGAGTCACAGAGATTTTAC	43

EF1α, elongation factor 1α.

Radioimmunoassay

The levels of tPRL₁₈₈ and tPRL₁₇₇ in the collected media samples were measured by homologous RIA (44, 45), using primary antibodies (antisera) raised in rabbit against tPRL₁₈₈ and tPRL₁₇₇ (anti-tPRL₁₈₈ and anti-tPRL₁₇₇, respectively) and secondary antibody raised in goat against rabbit IgG (anti-rabbit IgG; Sigma-Aldrich). Briefly, 50 μL of normal rabbit serum-EDTA-PBS (1% [vol/vol] normal rabbit serum, 0.05M EDTA, and 0.01M PBS) containing primary antibody was mixed with 50 μL of samples in RIA buffer, followed by 50 μL of labeled hormone in RIA buffer; after incubation at 4°C overnight, 100 μL of EDTA-PBS containing secondary antibody and 10% (wt/vol) polyethylene glycol was added to precipitate the antibody-bound hormone, as previously described (44). All antibodies used and their dilutions are listed in Table 2. Our cross-reactivity tests confirmed that the antisera raised against tPRL₁₈₈ and tPRL₁₇₇ bind specifically to those hormones (Supplemental Figure 2). For experiments using tPRLs, all incubation media, whether containing tPRLs (0.01, 0.1, 1, or 10 μg/mL) or not, were subjected to RIA. The measured values of incubation media without tPRLs were subtracted from those with added tPRLs yielding a measure for tPRLs added to media. To avoid the codetection of tested and released hormones, the measured values of added tPRLs were subtracted from the measurements of corresponding samples. The results from static incubations are shown as relative percent change from hyperosmotic controls, which did not contain any added PRLs. For perifusion experiments, the measurements of each fraction are plotted as percent change from the mean of 5 baseline fractions, which represents the release of tPRLs in the first 25 minutes of each experiment, where the cells were perifused with hyperosmotic medium (0.96–2.68 ng/mL for tPRL₁₈₈ and 1.43–2.93 ng/mL for tPRL₁₇₇, respectively; see the explanation for P_base below).

Table 2.

Antibody Table

Peptide/Protein Target	Antigen Sequence (if Known)	Name of Antibody	Manufacturer, Catalog Number, and/or Name of Individual Providing the Antibody	Species Raised in; Monoclonal or Polyclonal	Dilution Used
tPRL188		Anti-tPRL188	E. Gordon Grau	Rabbit; polyclonal	1:35 000
tPRL177		Anti-tPRL177	E. Gordon Grau	Rabbit; polyclonal	1:15 000
Rabbit IgG		Anti-rabbit IgG	Sigma-Aldrich, R1131-2ML	Goat; polyclonal	1:100

Concentration of tPRLs in RPD interstitial fluid

The interstitial concentration of tPRLs in tilapia RPD was estimated based on: 1) the study by Dharmamba and Nishioka (36); 2) the weight of 5 RPDs of FW-acclimated fish; and 3) the baseline release of tPRLs in perifusion experiments (Supplemental Tables 1–3). First, the volume of a single tPRL cell of FW-acclimated fish was calculated (Supplemental Table 1). Dharmamba and Nishioka (36) counted 105 cells per 6400 μm² (square of measurement) of cross-sectioned RPD of FW-acclimated fish. Based on the estimation that interstitial fluid accounts for 15% of tissue weight (46) and that the relative density of interstitial fluid and cells is 1.0 g/cm³, tPRL cells were estimated to occupy 85% of the measured area (cell area/square; 5440 μm²). The cross-sectional area of a single tPRL cell (S_c; 51.81 μm²) was obtained by dividing the cell area (5440 μm²) by the number of tPRL cells (FW tPRL cells/square; 105 cells). Assuming tPRL cells are spherically shaped, the radius (R_c; 4.06 μm) and volume (V_c; 280.33 μm³) of tPRL cells were calculated as follows:

$$\begin{array}{c}{R}_c=\sqrt{S_c/\pi }\\ V_c=4\pi {\left(R_c\right)}^3/3\end{array}$$

Then, the cell volume (cell volume, μm³; the volume that tPRL cells occupy in the RPD) and number of tPRL cells per RPD (RPD_c), as well as the volume of RPD interstitial fluid (RPD_i, μL), were calculated for 5 RPDs (mg) collected from FW-acclimated tilapia (144–217 g) based on the following equation (Supplemental Table 2):

$$\begin{array}{c}\text{Cell}\text{volume}=\text{RPD}\text{weight}\times 0.85\times {10}^9\\ RP{D}_c=\text{Cell}\text{volume}/V_c\\ RP{D}_i=\text{RPD}\text{weight}\times 0.15\end{array}$$

Means of calculated RPD_i (RPD_im; 0.0693 μL) and RPD_c (RPD_cm; 1 400 849 cells) were considered to represent the averages of RPD interstitial fluid volume and PRL cell number per RPD of fish employed in the present study. These values were used in further calculations.

The baseline concentration of tPRLs in RPD interstitial fluid (P_rpd, μg/mL = ng/μL) was calculated according to the baseline release of tPRLs in 10 experimental runs, totaling 54 chambers as shown below.

$$P_{rpd}=\left(P_{base}\times V\right)/\left\{\left(N_c/RP{D}_{cm}\right)\times RP{D}_{im}\right\}$$

Where P_base, baseline release of tPRLs (ng/mL; mean of 5 baseline measurements in each experiment); V, chamber volume (mL); and N_c, number of tPRL cells per chamber. These parameters were calculated as averages of 3–8 chambers in a single run (Supplemental Table 3). The numerator (P_base × V) represents the net release of tPRLs per chamber (ng). In the denominator, N_c/RPD_cm represents the relative number of PRL cells plated per chamber from an average-sized RPD (RPD_cm). By multiplying N_c/RPD_cm by the volume of RPD interstitial fluid corresponding to RPD_cm (RPD_im), the denominator represents the estimated volume of RPD interstitial fluid (μL) corresponding to the number of tPRL cells per chamber (N_c). This calculation assumes a uniform distribution of tPRL molecules in the RPD.

Statistics

Data were analyzed by two-way ANOVA with medium osmolality and concentration of tested PRLs as main effects in static incubations, and time course and treatment as main effects in perifusion experiments. Significant effects were followed up by Fisher's least significant difference (LSD) test. When necessary, data were log transformed to satisfy normality and homogeneity of variance requirements. All statistics were performed using Prism 6 (GraphPad), and data are shown as mean ± SEM.

Results

Static incubation with oPRL

Dose-related effects of oPRL on the release and gene expression of tPRLs under hyper- or hyposmotic conditions were examined employing a static incubation system (Figure 1). A two-way ANOVA detected significant individual and interaction effects of medium osmolality and oPRL concentration on the release of both tPRL₁₈₈ and tPRL₁₇₇ (Figure 1, A and B). Hyposmotically induced PRL release was observed for both tPRL isoforms at all oPRL doses tested. The release of both tPRL₁₈₈ (Figure 1A) and tPRL₁₇₇ (Figure 1B) was stimulated significantly by the two highest concentrations of oPRL (1 and 10 μg/μL) in hyposmotic medium. In hyperosmotic medium, 0.1-μg/mL oPRL suppressed the release of tPRL₁₇₇ (Figure 1B).

Figure 1.

Dose-related effects of oPRL on the release (A and B) and gene expression (C and D) of tPRLs from the dispersed tPRL cells in static incubation. A and C, tPRL₁₈₈. B and D, tPRL₁₇₇. Data are expressed as percent change from controls (0 μg/mL in 355 mOsm/kg) ± SEM. *, **, and ***, significant from 0-μg/mL controls in each osmolality at P < .05, P < .01, or P < .001, respectively; †† and †††, significant from hyperosmotic medium at P < .01 or P < .001, respectively. Two-way ANOVA followed by Fisher's LSD test.

The significant individual and interaction effects of medium osmolality and oPRL concentration were also detected on the mRNA expression of both tPRLs (Figure 1, C and D). Unlike release, gene expression of tPRLs was not altered by medium osmolality alone. In hyposmotic medium, oPRL suppressed the gene expression of tPRL₁₈₈ (Figure 1C) and tPRL₁₇₇ (Figure 1D) at the same concentrations that stimulated the release of tPRLs (1 and 10 μg/μL); tPRL mRNA expression in these groups was significantly lower than that in hyperosmotic groups.

Perifusion with oPRL

tPRL cells were perifused with either hyper- or hyposmotic media and with or without oPRL at the highest concentration examined in static incubation (Figures 2 and 3). Significant individual and interaction effects of time course and treatment were found on the release of both PRLs. Regardless of extracellular osmolality, oPRL significantly enhanced the release of both tPRLs during the entire course of exposure; tPRL₁₇₇ (Figures 2B and 3B) was stimulated to a greater extent than tPRL₁₈₈ (Figures 2A and 3A). The release of both tPRLs returned to baseline levels by 5–10 minutes after removal of oPRL from the medium. Both experimental groups responded stereotypically to hyposmotic stimuli either after or before the exposure to oPRLs by exhibiting transient spikes in the release of tPRLs (1037%–2430% of baseline).

Figure 2.

Effects of oPRL on the release of tPRLs from tPRL cells perfused in hyperosmotic medium. A, tPRL₁₈₈ release. B, tPRL₁₇₇ release. Data are expressed as percent change from baseline release ± SEM. White and black circles represent control and oPRL groups, respectively. *, significantly different from control at P < .05. Two-way ANOVA followed by Fisher's LSD test.

Figure 3.

Effects of oPRL on the release of tPRLs from tPRL cells perfused in hyposmotic medium. A, tPRL₁₈₈ release. B, tPRL₁₇₇ release. Data are expressed as percent change from baseline release ± SEM. White and black circles represent control and oPRL groups, respectively. *, significantly different from control at P < .05. Two-way ANOVA followed by Fisher's LSD test.

Static incubation with tPRLs

The dose-related and osmotically dependent effects of tPRL₁₈₈ and PRL₁₇₇ on their own release and gene expression were tested at concentrations employed in our previous study using gill filament (0.01–1 μg/mL) (Figure 4) (47) and at an additional high concentration (10 μg/mL) (Figure 5). In Figure 4, A and B, a two-way ANOVA revealed a significant individual effect of medium osmolality on the release of both tPRLs, in the presence or absence of added tPRLs. A significant individual effect of tested tPRL₁₈₈ concentration was found for both tPRL₁₈₈ and tPRL₁₇₇ release, whereas the effect of tPRL₁₇₇ concentration was only significant for tPRL₁₈₈ release. Employed as a positive control, hyposmotically induced PRL release in the absence of added tPRLs was confirmed for both tPRL isoforms (Figure 4, A and B). Regardless of extracellular osmolality, 0.01- and 1-μg/mL tPRL₁₈₈ suppressed the release of both tPRL₁₈₈ (Figure 4A) and tPRL₁₇₇ (Figure 4B); additionally, in both hyper- and hyposmotic media 0.1-μg/mL tPRL₁₈₈ suppressed its own release (Figure 4A). When added to hyposmotic medium, 0.01-μg/mL tPRL₁₇₇ suppressed tPRL₁₈₈ release, whereas in hyperosmotic medium, it stimulated tPRL₁₈₈ release at 1 μg/mL (Figure 4A).

Figure 4.

Dose-related effects of tPRLs (0.01–1 μg/mL) on the release (A and B) and gene expression (C and D) of their own from the dispersed tPRL cells in static incubation. A and C, tPRL₁₈₈. B and D, tPRL₁₇₇. Data are expressed as percent change from controls (0 μg/mL in 355 mOsm/kg) ± SEM. *, **, and ***, significant from 0 μg/mL controls in each osmolality at P < .05, P < .01, or P < .001, respectively; †, ††, and †††, significant from hyperosmotic medium at P < .05, P < .01, or P < .001, respectively. Two-way ANOVA followed by Fisher's LSD test.

Figure 5.

Effects of high-dose tPRLs (10 μg/mL) on the release (A and B) and gene expression (C and D) of their own from the dispersed tPRL cells in static incubation. A and C, tPRL₁₈₈. B and D, tPRL₁₇₇. Data are expressed as percent change from controls (0 μg/mL in 355 mOsm/kg) ± SEM. ** and ***, significant from 0 μg/mL controls in each osmolality at P < .01 or P < .001, respectively; †† and †††, significant from hyperosmotic medium at P < .01 or P < .001, respectively. Two-way ANOVA followed by Fisher's LSD test.

A significant individual effect of medium osmolality was detected on tPRL₁₈₈ gene expression only in the presence of added tPRL₁₈₈ (Figure 4C). A significant individual effect of tested tPRL₁₇₇ concentration was found for both tPRL₁₈₈ and tPRL₁₇₇ mRNA levels (Figure 4, C and D). Both added tPRL₁₈₈ and tPRL₁₇₇ showed an interaction effect with medium osmolality on their own mRNA levels. Consistent with the results from the experiment using oPRL, medium osmolality did not affect gene expression of tPRLs in the absence of tested tPRLs. Addition of tPRL₁₈₈ to hyposmotic medium stimulated gene expression of tPRL₁₈₈ at 0.1 μg/mL (Figure 4C), without altering that of tPRL₁₇₇ (Figure 4D). On the other hand, tPRL₁₇₇ suppressed gene expression of both tPRL₁₈₈ (0.01 and 1 μg/mL in hyperosmotic medium, and 0.01, 0.1, and 1 μg/mL in hyposmotic medium) (Figure 4C) and tPRL₁₇₇ (1 μg/mL in hyperosmotic medium, and 0.01, 0.1, and 1 μg/mL in hyposmotic medium) (Figure 4D). A significant difference between 355- and 300-mOsm/kg media was observed only for tPRL₁₈₈ mRNA with 1 μg/mL of added tPRL₁₈₈ (Figure 4C).

When the effects of tPRLs were tested at a higher concentration (10 μg/mL) (Figure 5), significant individual effects of medium osmolality and concentration of tPRLs (0 or 10 μg/mL) were observed on the release of both tPRL₁₈₈ and tPRL₁₇₇. An interaction effect between these factors was detected on the release of tPRL₁₈₈ regardless of tPRLs tested, and on tPRL₁₇₇ release only with the addition of tPRL₁₇₇. The release of tPRL₁₈₈ was stimulated by tPRL₁₈₈ in both hyper- and hyposmotic media, and by tPRL₁₇₇ only in hyposmotic medium (Figure 5A). This pattern was reversed for tPRL₁₇₇ release, where the stimulation by added tPRL₁₇₇ was observed regardless of medium osmolality, and that by tPRL₁₈₈ was detected only in hyposmotic medium (Figure 5B). Hyposmotically induced release of tPRLs was observed both in the presence and absence of added tPRLs (Figure 5, A and B).

Unlike release, neither medium osmolality nor added tPRLs, individually, were effective at eliciting the gene expression of tPRLs. An interaction effect was observed for tPRL₁₇₇ mRNA in the presence of added tPRL₁₈₈, which suppressed tPRL₁₇₇ mRNA in hyperosmotic medium (Figure 5, C and D).

Perifusion with tPRLs

tPRL cells were perifused with either hyper- or hyposmotic media and with or without 1-μg/mL tPRLs (Figures 6 and 7). Significant individual and interaction effects of time course and treatment were found on the release of both PRLs. Under hyperosmotic conditions, both tPRLs significantly enhanced the release of their own and, to a lesser extent, of one another (Figure 6). In hyposmotic medium, the stimulatory effect of tPRL₁₈₈ on its own release was limited to 2 out of 13 fractions; by contrast, tPRL₁₈₈ release was significantly enhanced by tPRL₁₇₇ during the entire course of exposure (Figure 7A). Both tPRLs stimulated tPRL₁₇₇ release, but the effect of tPRL₁₇₇ was more prominent than that of tPRL₁₈₈ (Figure 7B). In hyposmotic medium, the magnitude of stimulation by tPRLs was greater for tPRL₁₇₇ release than for tPRL₁₈₈ release (Figure 7, A and B). All experimental groups exhibited transient spikes of tPRLs (566%–2266% of baseline) upon exposure to hyposmotic medium either before or after exposure to tPRLs.

Figure 6.

Effects of tPRLs on the release of tPRLs from tPRL cells perfused in hyperosmotic medium. A, tPRL₁₈₈ release. B, tPRL₁₇₇ release. Data are expressed as percent change from baseline release ± SEM. White, gray, and black circles represent control, tPRL₁₈₈ and tPRL₁₇₇ groups, respectively. * and †, significantly different from control at P < .05; asterisks and daggers denote significance of tPRL₁₈₈ and tPRL₁₇₇ groups, respectively. Two-way ANOVA followed by Fisher's LSD test.

Figure 7.

Effects of tPRLs on the release of tPRLs from tPRL cells perfused in hyposmotic medium. A, tPRL₁₈₈ release. B, tPRL₁₇₇ release. Data are expressed as percent change from baseline release ± SEM. White, gray, and black circles represent control, tPRL₁₈₈ and tPRL₁₇₇ groups, respectively. * and †, significantly different from control at P < .05; asterisks and daggers denote significance of tPRL₁₈₈ and tPRL₁₇₇ groups, respectively. Two-way ANOVA followed by Fisher's LSD test.

Estimation of interstitial tPRL concentration in the RPD

The baseline concentrations of tPRLs in RPD were calculated as described, based on the weight of 5 RPDs of FW fish and baseline tPRL release in perifusion experiments (Supplemental Tables 1–3). The estimated concentration of interstitial tPRL₁₈₈ and tPRL₁₇₇ range between 3.71–21.90 and 6.30–21.69 μg/mL, respectively, ensuring that the concentration of tPRLs applied in perifusion experiments (1 μg/mL) was within the physiological range (Supplemental Table 3).

Discussion

Autocrine regulation of PRL and other pituitary hormones has remained poorly understood largely due to the lack of suitable model systems that allow for the primary incubation of homogenous cell populations prepared from healthy, normal pituitaries. Here, we reveal that both oPRL and native tPRLs, the latter at physiological concentrations, stimulate the release of tPRLs from dispersed tPRL cells in a perifusion system in vitro. The onset of PRL-induced PRL release took place within 5–10 minutes of exposure and lasted the entire course of exposure. The effect subsided within 5–10 minutes after removal of PRLs from media. The two isoforms of tPRL exhibited differing abilities to stimulate both their own release and that of one another. The release of tPRL₁₈₈ and tPRL₁₇₇ responded differentially to the same hormonal stimulus. Moreover, the effects of tested PRLs differed in hyper- and hyposmotic media. These observations suggest that tPRL cells are directly regulated by the hormones they secrete through a positive feedback loop, which is in turn modulated by extracellular osmolality.

Previous studies have provided ample evidence of direct osmotic control of tPRL cell function (32, 34, 39, 48–50). We have also obtained evidence that the regulation of tPRLs' secretion by extracellular osmolality is modulated by a variety of endocrine factors, such as PRL-releasing peptide (51), somatostatin (52–55), GnRH (51, 56, 57), thyrotropin-releasing hormone (55), vasoactive intestinal peptide (58), urotensin II (53), dopamine (50, 55), angiotensin II (59, 60), 17-β estradiol (61, 62), testosterone (62), cortisol (54, 55, 62, 63), atrial natriuretic peptide (64), brain natriuretic peptide (64), ghrelin (65, 66), leptin (67), IGF-1 (68, 69) and IGF-2 (69), and ouabain (70) (for review, see Ref. 29). In the current study, we have discovered that the osmotic regulation of tPRL cells is also affected by the hormones they secrete. Additionally, the present study is the first to directly show that extracellular osmolality can modulate the autocrine regulation of hormone release. In static incubations in hyposmotic medium, oPRL further augmented the release of both tPRLs relative to hyposmotic medium alone, an effect that was not observed when cells were incubated in hyperosmotic medium (Figure 1). Moreover, both oPRL and tPRL₁₇₇ dramatically increased tPRL₁₈₈ release in hyposmotic perifusion medium (up to 801% and 867% of baseline, respectively) (Figures 3A and 7A), and attenuated stimulation in hyperosmotic medium (up to 269% and 640% of baseline, respectively) (Figures 2A and 6A). tPRL₁₈₈ also stimulated its own release, but only in hyperosmotic medium. These findings indicate that the interaction between the endocrine system and surrounding environments can occur at the level of single hormone-secreting cells via an autocrine feedback loop, without exogenous endocrine regulators. The multilayered-regulation of tPRL cells by hypothalamic and extrahypothalamic hormones, tPRLs and extracellular osmolality indicates that the regulation of PRL and possibly other pituitary hormones is more complex and nuanced than once believed.

Critical aspects of our experimental model were: 1) the employment of normal cells, instead of cancer cells; and 2) the use of a perifusion system to minimize the risk of confounding the actions of tested and released PRLs, which pose a clear limitation in the interpretation of static-incubation data. By estimating the basal concentration of interstitial tPRLs within the RPD, we demonstrated that the concentration of tPRLs we employed to stimulate tPRL cells (1 μg/mL) is well within the physiological range (Supplemental Tables 1–3). As far as we know, the current study is also the first direct demonstration of positive autocrine feedback of the release of a vertebrate pituitary hormone. By contrast, release of PRL from dispersed pituitary cells of grass carp was suppressed by oPRL in a dose-related manner (71). oPRL also suppresses transcription and secretion of PRL from cultured rat GH3, GC, and 2B8 cells, suggesting negative feedback regulation for PRL (14–16). When interpreting these conflicting results, it is important to consider whether tested PRLs have been previously applied within their physiological concentrations. It is difficult to answer this question, inasmuch as descriptions on the estimated interstitial PRL concentration within the pituitary are rarely reported. Nonetheless, Herbert et al (15) observed the negative regulation of endogenous PRL release by oPRL at lower concentrations (0.01–10 ng/mL) than the basal level of circulating PRL in humans and other mammalian species (up to 25 ng/mL) (72). Inasmuch as the concentrations of PRLs are physiological, the differences in autocrine regulation between tPRLs and other species' PRLs suggest they may play distinct roles in the pituitary and other target tissues.

Negative feedback regulation is essential for maintaining stable levels of most hormones in circulation. On the other hand, examples of positive feedback in nonhealthy/normal tissues are limited (see Ref. 73). Both RPDs and PRL cells of tilapia in FW are larger than those of fish in SW (36), suggesting that the growth and proliferation of PRL cells are stimulated in hyposmotic conditions. Considering the hyperosmoregulatory actions of PRL, it is conceivable that the observed positive autocrine loop serves to bolster the production of PRL during the critical acclimation from SW to FW, when the fish needs it most. This may explain why our findings differ from what has been observed in mammalian models, in which PRL has been proposed to act as an antiproliferative and/or proapoptotic factor in anterior pituitary cells, including PRL cells (19, 20). It is important to point out, however, that the patterns of gene expression of both tPRLs did not generally follow those of release; although release under static incubations was under positive autocrine regulation, gene expression was either unresponsive or suppressed by added PRLs in most cases. These observations suggest that the autocrine stimulation of release of both tPRLs and its modulation by osmolality may be short-lived.

In the current study, the two isoforms of tPRL elicited different patterns of autocrine regulation. Previous studies have shown that tPRL₁₇₇ and tPRL₁₈₈ differ in the magnitude of their responses to environmental osmolality in vivo (32, 45, 74, 75) and in vitro (32, 75), and in their actions (74, 76). This could be partially attributed to an isoform-specific regulation of their gene expression (32). Differences in posttranscriptional regulation of tPRLs were also suggested (74), but this question has not been further addressed after Specker et al (77) reported the colocalization of tPRL₁₈₈ and tPRL₁₇₇ proteins within the same secretory granules in tPRL cells. The acute onset and termination of PRL-induced PRL release suggest that the tested PRLs triggered the release of available tPRLs from intracellular stores. The isoform-specific differences in this autocrine response further suggest that the regulatory mechanisms underlying PRL-induced tPRL₁₈₈ and tPRL₁₇₇ release are distinct. Although signal transduction mechanisms involved in rapid hyposmotically induced PRL release (33, 78–81) and PRL actions on target cells (82) have been examined separately, the manner in which these mechanisms interact in the regulation of PRL cells should be addressed in future studies.

Taken together, the present results employing tPRL cells revealed a novel aspect of the vertebrate endocrine system, consisting of positive autocrine feedback that is modulated by extracellular osmolality. This adds to the evidence that the regulation of PRL is multifactorial. Although the study of the regulation of PRL cell function by individual factors in isolation offers valuable information, it is now clear that a full understanding can only be obtained when the many levels of regulation and their interactions are considered comprehensively.

Abbreviations

FW

freshwater

LSD

least significant difference

oPRL

ovine PRL

PRL

prolactin

PRLR

PRL receptor

qRT-PCR

quantitative real-time PCR

RPD

rostral pars distalis

SW

seawater

TFA

trifluoroacetic acid

tPRL

tilapia PRL.