| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLE |
-Subunit Genes by Gonadotropin-Releasing Hormone and Estradiol-17ß in Single Rat Pituitary Cells during Aggregate Cell Culture
Laboratory of Cell Pharmacology, University of Leuven (K.U. Leuven), Medical School, Campus Gasthuisberg (O&N), B-3000 Leuven, Belgium
Address all correspondence and requests for reprints to: Professor Carl Denef, Laboratory of Cell Pharmacology, University of Leuven, Medical School, Campus Gasthuisberg (O&N), B-3000 Leuven, Belgium. E-mail: carl.denef{at}med.kuleuven.ac.be.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
mRNA together with other hormone mRNAs. GnRH strongly increased the proportion of cells containing GSU mRNA alone, but E2 did not. GnRH and (E2) affected the expansion of a population (
20% of all cells) coexpressing PRL and GSU mRNA without ßGSUs. Immunostaining of stored hormone on tissue sections revealed colocalization of PRL and GSU in the E2- but not in the GnRH-treated cells. The present findings suggest that cells coexpressing different pituitary hormone mRNAs form a distinct population that survives without extrapituitary factors. Their occurrence can be markedly modified by regulatory factors. Certain hormone regimens favor unique coexpressions distinctly at mRNA and protein level. These peculiar characteristics support the notion that combinatorial expression of hormone genes in the pituitary serves a biological role. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The function and regulatory control of bihormonal cells remains unclear. Some investigators have described the presence of bihormonal cells in several physiological and pathological conditions characterized by a high demand of one of the hormones expressed by the bihormonal cells. Therefore, it was proposed that bihormonal cells might be an intermediate stage of monohormonal cells shifting from one hormone type to another (11). However, bihormonal cells have also been observed under basal conditions in which hypersecretion of a specific hormone is not required (13, 14).
We previously showed, by means of single-cell RT-PCR, that a considerable number of rat and mouse anterior pituitary cells contain two or more hormone mRNAs. Indirect evidence was found that not all these cells also store the respective hormones in a detectable amount, as the classical monohormonal pituitary cell types do (15). On the basis of apparent discrepancy between expression at the mRNA level and at the protein level, the latter cell population was proposed to be a putative subpopulation of pituitary cells with a putative functional significance. Additional evidence for a putative biological significance came from the findings that the pattern of coexpressed hormone mRNAs differs significantly according to age, although cells plurihormonal at the mRNA level are already present at d 16 of fetal development in mice (16). Moreover, as judged from intracellular free calcium responses, certain regulatory peptides preferentially target cells displaying combinatorial mRNA expression (15).
The present investigation was intended to gather further evidence for a biological significance of combinatorial hormone mRNA expression. Investigating whether cells expressing a particular combination of hormone mRNAs increase (or regress) in number on chronic exposure to certain regulatory factors may give us this evidence. The classical gonadotrophs, for example, increase significantly in number and size on removal of the gonads because of chronic interruption of the negative feedback by gonadal hormones, resulting in gonadotroph proliferation in a GnRH-dependent manner (17, 18, 19), and classical lactotrophs develop hyperplasia during chronic treatment with estrogen (20, 21). Evaluating the stimulatory effects of GnRH and estradiol E2 on combinatorial hormone mRNA expression seems therefore helpful in our search for a putative biological relevance of these cells.
We evaluated whether pituitary cells expressing combinations of hormone mRNAs persist on removing them from the in vivo regulatory machinery by studying whether these cells survive after 4 wk in culture. On the other hand, we investigated whether their number in culture changes on chronic (4 wk) exposure in vitro to regulatory agonists (GnRH and E2) known to affect the population size of specific classical cell types. We also evaluated whether those changes would be similar or dissimilar to those in the classical monohormonal cells. A reaggregate cell culture system was used because in this type of culture, pituitary cells remain differentiated and functional for at least 2 months (22, 23, 24, 25, 26).
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following primary antisera were used for immunostaining: rabbit antirat PRL (anti-rPRL-IC-5) and guinea pig antirat glycoprotein hormone -subunit (GSU) obtained from Dr. A. F. Parlow through the National Hormone and Pituitary Program (NIDDK, Harbor-UCLA Medical Center, Torrance, CA). Furthermore, normal swine serum, normal goat serum, normal donkey serum, and avidin biotinylated complex/horseradish peroxidase were obtained from DAKO Corp. (Glostrup, Denmark); secondary antibodies goat antirabbit fluorescein isothiocyanate (FITC)-labeled and donkey antiguinea pig Texas Red-labeled IgGs from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA); and Vectashield from Vector Laboratories (Burlingame, CA). 3-Amino-9-ethylcarbazole tablets, H2O2, bovine pancreatic trypsin type III, and porcine pancreatic trypsin were purchased from Sigma Aldrich; paraformaldehyde from Riedel-deHaën (Seelze, Germany); and picric acid from UCB (Braine-l’Alleud, Belgium). Paraffin was obtained from Klinipath (Turnhout, Belgium).
Animals
Fourteen-day-old female Wistar rats were obtained from the University Animal Breeding Facility (Heverlee, Belgium). Mice (FVB/NhanHsd) were purchased from Harlan (Horst, The Netherlands). All animals were kept in an environment of constant temperature, humidity, and day-night cycle with free access to water and food within the animal house facilities (University of Leuven, Belgium). All animal experiments were conducted in accord with the Guidelines for Care and Use of Experimental Animals and were approved by the University Ethical Committee. Rats were killed by decapitation and mice by CO2 gassing.
Preparation of single pituitary cells and reaggregate cell cultures
Pituitaries were dispersed into single cells and either applied onto a FACS to obtain one cell per well (see below) or established in reaggregate cell culture by methods described previously (22). Briefly, pituitaries were treated sequentially with 2.5% porcine pancreas trypsin, deoxyribonuclease (DNase), trypsin inhibitor, EDTA in Ca2+- and Mg2+-free medium after which they were mechanically dispersed through a narrowed Pasteur pipette. This was followed by a second DNase-treatment step after which debris was removed by centrifugation through a 3% BSA layer. The cells were seeded in 35-mm nontreated polystyrene dishes in 2 ml defined culture medium at 2 x 106 cells per dish. The culture medium was a serum-free defined medium made by Life Technologies, Inc. (Invitrogen) as described previously (25). Shortly it consisted of DMEM/F12 1/1, supplemented with 0.6 µM iron (Fe2+ and Fe3+), 5 mg/liter insulin-Zn, 5 mg/liter transferrin, 50 mg/liter streptomycin, 35 mg/liter penicillin, 10 mM ethanol, 1 mg/liter catalase, and 1 g/liter NaHCO3. Cells were allowed to reaggregate on a gyratory shaker at 63 rpm in a 1.5% CO2 water-saturated incubator at 37 C (25).
The latter in vitro system has been extensively used in our laboratory over the last two decades for in vitro studies on secretion, proliferation, and differentiation of anterior pituitary cells and has been shown to be superior to monolayer cell culture at the structural, ultrastructural, functional, and developmental level (22, 23, 24, 25, 26). For example, hormone secretion is highly sensitive to all classical releasing and inhibiting hormones (22), typical cell type associations occur as seen in the normal pituitary (27), the three-dimensional nature allows paracrine interactions to occur (28), and the classical cell types are retained in normal proportions for up to 2 months in culture (29) without evidence for fibroblast proliferation (26).
For coaggregate cell cultures, 1 x 106 pituitary cells from 14-d-old female rats were mixed with 1 x 106 pituitary cells from adult male mice per dish.
Treatment of reaggregate pituitary cell cultures with GnRH or E2
After 3 d of culture, aggregates were pooled and subsequently transferred to 55-mm dishes in 6 ml fresh defined culture medium at an equivalent of 2 x 106 cells per dish. In every experiment at least three separate dishes were treated with GnRH or E2, and no substance was added to another three or more dishes (control condition). GnRH or E2 was added at a concentration of 1 nM. To control dishes, no substances were added because the vehicle components (BSA for GnRH and ethanol for E2) are already present at a higher concentration in the defined medium (22). The medium was renewed twice a week. Each time E2 and GnRH were added simultaneously with the fresh medium. A dose of 1 nM GnRH was used, which is a submaximal concentration but above the physiological concentrations in terms of LH secretory responses (30). The dose of E2 was also 1 nM and was previously shown to significantly increase the number of cells expressing PRL mRNA in aggregate cell cultures as detected by in situ hybridization (31). These supraphysiological doses were used to ensure cell population expansion of sufficient magnitude and, hence, statistical power because the labor-intensive nature of the single-cell RT-PCR approach set certain limits to the number of cells that can be analyzed.
Redispersion of reaggregates and isolation of single cells by FACS
After 4 wk in culture, all aggregates from dishes belonging to the same experimental condition (E2 treatment, GnRH treatment, or control) were pooled and dispersed using an analogous procedure as for the preparation of single cells. Aggregates were treated sequentially with 2.5% porcine pancreas trypsin, trypsin inhibitor, and EDTA in Ca2+- and Mg2+-free medium, followed by mechanical dispersion through a Pasteur pipette. DNase treatment was performed and debris was removed through a 3% BSA layer. After redispersion, cells were suspended in FACSFlow sheath fluid at 1 x 106 cells per milliliter and kept on ice until sorting.
Single cells were deposited at one cell per well with a FACS Vantage equipped with a one-cell deposition unit in 96-well PCR plates (Thermofast 96) already containing 5 µl lysis buffer (0.05% IGEPAL CA-630, 4 U ribonuclease inhibitor, 2 mM dithiothreitol, 2x first-strand buffer). FACS parameters were optimized so that not more than one cell was deposited per well. FACS settings ensured that cell selection was random, as was confirmed by comparable scatter (FSC-SSC) profiles of the cell population before and after FACS sorting (i.e. sorted under single-cell sorting conditions). Plates were centrifuged for 5 min at 3000 rpm and 4 C to bring both lysis buffer and cell together at the bottom of the well. Cells were lysed during a 30-min incubation step on ice and snap frozen by immersing the tubes in a mixture of ethanol and dry ice. Cell lysates were stored in a -70 C freezer until further use.
Single-cell RT-PCR
Reverse transcription (RT).
RT was performed on total-cell lysate (5 µl). The samples were heated for 10 min at 70 C to disrupt secondary structures. After quick cooling to 4 C, 5 µl RT mixture, containing 5 µM random hexamers, 10 U Moloney murine leukemia virus reverse transcriptase, and 0.5 mM of each dNTP, was added. The samples were thereafter kept for 15 min at 25 C followed by 50 min at 42 C and 10 min at 95 C. A positive control was included by adding 100 pg total RNA from 14-d-old rat pituitary to one well. Total pituitary RNA extraction was performed with Tripure reagent.
PCR.
A 2-µl aliquot of the 10-µl RT reaction sample was added to 8 µl PCR mixture and subjected to PCR. Ingredients of the PCR mixture were AmpliTaq Gold DNA polymerase (0.75 U), dNTPs (0.1875 mM each), 3.125 mM MgCl2, 0.875 x PCR Gold buffer, and 1.25 µM specific sense and antisense primers. The primers are presented in Table 1
. PCR was performed for the following hormone cDNAs: GH, PRL, proopiomelanocortin (POMC), GSU, LHß, TSHß, and FSHß. The presence of a cell was checked in every sample using a PCR for L19, a rat ribosomal protein present in every cell. GH and PRL cDNA as well as the cDNA of TSHß and LHß were analyzed in the same (duplex) PCR. All amplification reactions started with a 10-min, 95 C step to activate the AmpliTaq Gold DNA polymerase (hot start). Then the samples were submitted to 45 cycles of three different temperatures: 10 sec at 95 C, 20 sec at the annealing temperature of 60 C (except for GSU for which an annealing step at 55 C was used), followed by an elongation step of 30 sec at 72 C. As an exception, optimization of the POMC PCR resulted in a cycling protocol of 10 cycles of 20 sec at 95 C, 30 sec at 45 C, and 30 sec at 72 C, followed by another 35 cycles of 10 sec at 95 C, 30 sec at 55 C, and 30 sec at 72 C. At the end of all reactions, an extra step of 7 min at 72 C was performed. Positive PCR controls consisted of 100 pg of reverse-transcribed rat total pituitary RNA. Two microliters PCR mixture was used instead of sample as a negative control. PCR was performed in a GeneAmp PCR system 9700.
|
Examples of RT-PCR products obtained from single-cell lysates are shown in Fig. 1.
|
) to amplify the whole coding sequence. Cloning of the POMC coding sequence has previously been described (15). Rat pituitary RNA was extracted with Tripure reagent and treated with DNase to exclude amplification of the respective genomic DNAs. RT was performed on this material in the same way as described above, using 1.75 µM random hexamers as well as 1.75 µM oligo dT primers to assure transcription of the full-length mRNA. PCR was performed with Ready-to-Go PCR beads and included 3 min at 94 C; 40 cycles of 30 sec at 94 C, 30 sec at annealing temperature (between 55 and 60 C), and 1 min at 72 C; and finally 10 min at 72 C. The amplified fragment was loaded on a gel of 1% agarose and Tris-boric acid-EDTA and extracted with a Recochip (TaKaRa). The fragment was inserted into a TOPO2.1 vector and the ligation sample used for transformation of TOP10F' Escherichia coli cells. Cells were then plated on Luria Broth medium agar containing 100 mg/liter ampicillin (LBamp) and incubated at 37 C overnight. A few colonies were picked and incubated in liquid LBamp overnight in a 37 C shaker. Plasmid DNA was isolated using a plasmid minikit (QIAGEN). The insert (EcoRI fragment) was sequenced using the Thermo Sequenase fluorescent-labeled primer cycle sequencing kit and analyzed on an automated laser fluorescent DNA sequencer (both from Amersham Pharmacia Biotech).
|
GSU, LHß, FSHß, and PRL approximately 40 copies. The sensitivity for POMC has previously been determined (50 copies) (15).
Reproducibility was tested by performing four independent PCRs for the same hormone on a few single-cell samples. As far as tests for the GSU and GH/PRL PCR, an average reproducibility of 85% was found.
Specificity controls for the single-cell RT-PCR method
When performing RT-PCR on single cells obtained by FACS, several sources of contamination can lead to false-positive results: 1) deposition of more than one cell per well during FACS sorting; 2) presence of cell debris from another cell on the surface of the sorted cell caused by incomplete dispersion; 3) presence of RNA from cells lysed during preparation in the small amount of fluid (10 nl), deposited together with each cell in a well; and 4) sticking of RNA from cells lysed during preparation to the sorted cell. Several controls were performed to evaluate these possible sources of contamination. To minimize possible contamination of sorted cells by RNA from lysed cells present in the cell suspension fluid, the cell suspension was diluted 100 times before FACS sorting. Before filling each plate, 5–10 single cells were sorted on a 96-well plate lid and the droplets deposited were microscopically inspected to ascertain that not more than one cell was present in every droplet. An experiment in which sorted cells were transferred to a standard flat-bottom 96-well plate, stained with the fluorescent DNA-binding dye thiazol orange, and microscopically examined confirmed the presence of not more than one cell per well. Transmission electron microscopy (CM10, Philips, Eindhoven, The Netherlands) analysis of the cell suspension did not reveal that debris of cells was adhering to cell membranes of intact cells (data not shown).
In another experiment, cells redispersed from control aggregates and from aggregates treated with E2 for 2 wk in culture were spun down and the supernatant serially diluted and tested with RT-PCR for presence of the high-abundant PRL and GH mRNAs. It was found that supernatant in a 10-fold excess of the fluid present in a well with the single cell (10 nl) after FACS sorting did not yield a positive PCR signal. In a more rigorous test, aggregates consisting of a 1/1 mixture of pituitary cells from 14-d-old rat and adult mouse were prepared and maintained for 2 wk in culture. Mouse and rat pituitary cells are homogeneously mixed in such coaggregates as described before (16). Transmission electron microscopy analysis showed normal cell-cell contacts in the mixed aggregates. After redispersion and FACS sorting of the cells, RT-PCR was performed with fully selective primers for rat PRL mRNA and mouse GSU mRNA (Table 1
, Srat and Smouse). These primers allowed PCR amplifications with the same sensitivity as achieved by PCR with the standard primers, as tested on serial dilutions of total RNA extracted from 14-d-old rat and adult mouse pituitary. The FACS’s parameter settings ensured that the 1/1 distribution of rat and mouse cells was not significantly altered after sorting as confirmed by similar scatter-profiles (FSC vs. SSC) of the cell population before and after (single) cell sorting. Of the 160 cells tested, 16.3% of the cells presumed to be rat cells (i.e. 13/80 cells) contained rat PRL mRNA and 11.1% of the mouse cells contained mouse GSU mRNA. Colocalization of the two mRNAs was not detected, demonstrating the absence of cell debris or exogenous RNA sticking to the sorted cell. Moreover, cells in the original cell suspension were spun down (5 min at 1500 rpm and 4 C); the cell suspension medium was diluted to a same level as cells before sorting (1/100) and sorted in a 96-well plate containing lysis buffer, using electronic test pulses to obtain a same nl amount of the supernatant (without cell) in each well. When RT-PCR was performed subsequently, no PCR signals were found for rat PRL mRNA and mouse GSU mRNA, again excluding the presence of detectable mRNA levels from lysed cells in the suspension medium at the dilutions used.
Fluorescent and carbazole immunostaining
After 1 month in culture some aggregates were examined for the presence of PRL and GSU using methods described previously (32). Briefly, aggregates were fixed for 4 h in Zamboni solution (4% paraformaldehyde in 15% saturated picric acid solution and 0.1 M phosphate buffer, pH 7.4) and after dehydration embedded in paraffin. The 2-µm tissue sections were permeabilized by trypsinization (bovine pancreatic trypsin, 0.1%) for 5 min, preincubated for 1 h with preimmune serum (20% in Tris-buffered saline, pH 7.4) from the animal species in which the secondary antibodies had been raised, and incubated overnight at room temperature with the primary antibodies. Final dilutions were 1/5000 for rabbit antirat PRL-IC-5 and 1/1000 for guinea pig antirat GSU. For the fluorescent (double) staining, fluorochrome-labeled secondary antibodies (FITC-conjugated goat antirabbit 1/400 and Texas Red-conjugated donkey antiguinea pig 1/400) reacted with the sections for 1 h at room temperature while protected from light. Sections were mounted in Vectashield, and kept at -80 C. Sections were examined on a Zeiss Axioplan 2 microscope (Carl Zeiss, Weesp, The Netherlands) equipped with an appropriate filter (FITC/Rhod/DAPI), and coupled to a Spot Advanced camera (Carl Zeiss), version 2.2.2. The conventional immunostaining was performed with a biotinylated swine antirabbit secondary antibody in a 1/400 dilution for the PRL staining and a biotinylated donkey antiguinea pig antibody (dilution 1/500) for the GSU staining, followed by 1-h incubation with the avidin biotinylated complex/horseradish peroxidase. Thereafter carbazole was used for staining, i.e. precipitated by an oxidation reaction in combination with hydrogen peroxide and the peroxidase. Examination of these sections was performed using a light optic microscope (Orthoplan, Leitz, Wetzler, Germany) and a computer-image analysis system equipped with color detection (Quantimet 500; Leica Corp., Cambridge, UK). Stained area as a proportion of total section area was measured and the number of stained cells per 105 µm2 was counted for 30 fields on each slide. Two slides (from two independent experiments) were evaluated for each condition.
Omission of (one of) the primary or secondary antibodies did not result in staining; neither did replacing the primary antibody with nonimmune serum (normal rabbit serum diluted 1/5000 and normal guinea pig serum diluted 1/1000).
Experimental designs and statistical analysis
Statistical analysis was performed with the help of the Statistics Center of our faculty.
Test pattern.
Because of the limited amount of single-cell cDNA, it was not possible to amplify all pituitary hormone cDNAs by PCR for every cell. Therefore, three distinct combinations (called test pattern farther in the text) of hormone cDNA amplifications were performed on different sets of cells within each of the four experimental conditions (original cell population, control cultured cells, E2-treated cultured cells, and GnRH-treated cultured cells). In a first set, all hormone cDNAs were amplified except POMC cDNA, the second set included all hormone cDNAs except GSU cDNA, and in the third set, all hormone cDNAs were amplified except GH and PRL cDNA.
Cell culture experiments.
Five independent experiments were run. The first experiment consisted of a control condition, a GnRH treatment condition, and an E2 treatment condition. The next two experiments included a control and GnRH treatment group, and the last two a control and E2 treatment group. Hence, for each complete treatment set, three independent experiments were done, each time in combination with an untreated condition. A total amount of 769 cells belonging to the control condition, 507 cells belonging to the E2-treated condition, and 426 cells belonging to the GnRH-treated condition were analyzed.
Freshly dispersed pituitary cells.
Hormone mRNA distribution in the freshly dispersed pituitary cells was analyzed in four independent preparations (613 cells in total). These were, however, not obtained simultaneously with the preparations used for the cell cultures.
Statistical analysis.
In a first analysis, two comparisons were made: 1) the control condition vs. the freshly dispersed cells and 2) the control condition vs. GnRH- and E2-treated cells. These comparisons were made in two different ways. A cumulative logits model was used to compare the total population of hormone mRNA-expressing cells (cells not expressing any hormone vs. cells expressing one hormone vs. cells expressing more than one hormone). A multicategory logit model was used to compare the expression for each hormone separately (cells not expressing this hormone vs. cells expressing this hormone alone vs. cells expressing this hormone in combination with other hormones). To correct for differences in hormone mRNA expression because of possible differences between the experimental settings, experiment was added in the statistics model as a (fixed) covariate or random effect.
The statistical program performed a random selection of cells of the total number of cells analyzed to obtain almost identical distributions of the test pattern in each experimental condition (identical proportions of cells with the not measured hormone mRNA(s) in each experimental condition are obtained). As a consequence of this, the measured proportion of cells displaying a particular hormone mRNA expression is an underestimation of the real proportion of cells with this expression. This underestimation was similar in all groups and therefore does not burden the comparison between the different experimental conditions in the statistical analysis.
However, for illustrative purposes, the values used to represent the data in the figures were obtained by combining different sets of cells (having a different test pattern). The missing values in each individual set were added, based on the data obtained in another set from the same experimental condition. This quick and dirty method was used to get an idea of the real size of the different cell populations but could not be used for statistical analysis.
To compare the occurrence of cells expressing a specific combination of hormone mRNAs among the different experimental conditions, we included all cells in which these hormone mRNAs were measured. A logistic regression analysis was performed because there were no missing values (unlike in the previous statistical analysis).
Statistical analysis of immunostained hormone in sections of the pituitary aggregate cell cultures was as follows. Differences in stained areas and number of stained cells per unit section area were measured by computerized image analysis and were analyzed for statistical significance by ANOVA.
For clarity, cells expressing only one pituitary hormone mRNA will further be denoted as monohormonal cells and cells expressing more than one hormone mRNA as plurihormonal cells. For example, PRL monohormonal cells contain PRL mRNA only; PRL plurihormonal cells contain PRL mRNA together with other hormone mRNA(s) of any kind. Cells expressing both a ß-subunit of the glycoprotein hormones (ßGSU) and the GSU were taken as plurihormonal because the statistical analysis design used did not allow otherwise.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
30%) was not significantly different (P = 0.11) from the proportion of these cells in the cell suspension before culture (33.9%) (Fig. 2). The proportion of monohormonal cells also did not significantly change.
|
).
Occurrence of cells expressing a particular hormone mRNA alone or together with other hormone mRNAs
PRL mRNA-expressing cells.
Some changes in the proportion of pluri- and monohormonal cells expressing PRL mRNA were seen after culture (Fig. 3). However, none of these changes reached statistical significance. E2 treatment significantly increased both monohormonal and plurihormonal cells expressing PRL mRNA (Fig. 3) (P < 0.0001 for both increases). The magnitude of increase was similar for both cell groups (P = 0.78). Treatment with GnRH had a similar effect on plurihormonal cells expressing PRL mRNA (P = 0.0003). The expansion of the monohormonal PRL cell population was less pronounced but still significant (P = 0.042). The difference in magnitude of change between the mono- and plurihormonal population did not reach statistical significance (P = 0.068).
|
GSU mRNA-expressing cells.GSU mRNA were changed somewhat after culture (Fig. 4). However, none of these changes reached statistical significance. As shown in Fig. 4, E2 caused a small but significant increase in the proportion of plurihormonal cells expressing GSU (P = 0.03) but not in monohormonal cells expressing GSU. Addition of GnRH to the cultures resulted in a marked increase in both monohormonal and plurihormonal cells containing GSU mRNA (P < 0.0001 in both cases). There was no difference in magnitude of change between the plurihormonal and monohormonal population (P = 0.62).
|
). The fall in GH-expressing cell number likely is due to the absence of glucocorticoids in the culture medium because glucocorticoids are known to increase GH expression by a direct action at the pituitary level (33). E2 treatment caused a significant decrease in the proportion of monohormonal cells containing GH mRNA (P < 0.0001) but not in the proportion of plurihormonal cells expressing GH mRNA. GnRH tended to decrease the proportion of monohormonal and increase that of plurihormonal GH mRNA-expressing cells, but the changes were statistically not significant.
|
5.4%, 5.1%, and 3.2%, respectively) (data not shown). The proportion of monohormonal TSHß mRNA cells was increased from approximately 0.2% to approximately 2.5% (P = 0.013), but the proportion of plurihormonal TSHß cells was not (data not shown). In a few cells, one of the ßGSUs was found without the common GSU (data not shown). Neither GnRH nor E2 affected the proportional number of cells expressing POMC, LHß, FSHß, or TSHß mRNA alone or in combination with other hormone mRNAs (data not shown).
Effect of culture and GnRH or E2 treatment during culture on cells expressing particular combinations of hormone mRNAs
According to previous studies in our laboratory (15), combinatorial expression of hormone mRNAs is not a random phenomenon because cells responsive to a certain agonist (such as 
3-MSH) display a combined hormone mRNA expression that is markedly different from cells in the total pituitary cell population. Moreover, the prominence of certain combinations is age dependent (16). It was therefore important to investigate which particular hormone mRNA combinations are affected by deprivation from extrapituitary factors (culture) and treatment with GnRH and E2 in culture.
Freshly dispersed cells vs. cultured cells.
In accordance with our previous study (15), combined expression of PRL, GH, and TSHß mRNA (Pit-1-dependent hormone mRNAs) and of one of the latter hormone mRNAs together with mRNA of POMC or LHß/FSHß were observed. In addition, approximately 5% of the cultured cells expressed PRL mRNA together with GSU mRNA but without the mRNA of any of the ßGSUs or GH (Figure 6). In fact, before culturing, the latter cells were very scarce (0.6%).
|
). The proportion of cells expressing both PRL and GSU mRNA in the absence of ß-subunits or GH mRNA increased dramatically from 5% in the control condition to 18.3% after E2 treatment (P < 0.0001) and to 20.6% after GnRH treatment (P = 0.0024). In contrast, there was no significant effect in cells expressing PRL and GSU together with one of the ß-subunits of the glycoprotein hormones (LHß, FSHß, or TSHß) (control condition compared with GnRH treatment, E2 treatment and initial population gives, respectively, P = 0.84, P = 0.15, and P = 0.61). There was also no difference in the occurrence of cells coexpressing GH or POMC mRNA and GSU mRNA among the experimental groups: control, compared with GnRH treatment, E2 treatment, and initial population for GH+GSU mRNA gives, respectively, P = 0.62, P = 0.98, P = 0.07, and for POMC+GSU mRNA gives, respectively, P = 0.21, P = 0.96, P = 0.24. The proportion of cells with combinatorial expression of GH and PRL mRNA was not altered (P = 0.052, P = 0.58, and P = 0.78 for control, compared with GnRH treatment, E2 treatment, and initial population, respectively).
Immunostaining of expressed hormones in culture
To determine whether the effects of GnRH and E2 on the proportion of cells expressing PRL and GSU mRNA was also seen at the protein production level, immunostaining for PRL and GSU on sections of the cultured aggregates was done (two independent experiments). Both GSU- and PRL-immunoreactive cells were detectable after 1 month in culture. As shown in Table 3, there was a clear-cut differential effect of E2 and GnRH treatment. GnRH significantly increased the proportional area occupied by GSU immunoreactivity as well as the number of GSU-immunopositive cells (P < 0.0001). E2 treatment had no effect. An opposite effect was seen in case of PRL cells. E2 treatment significantly augmented the proportion of PRL-immunoreactive area (P < 0.0001) and the number of immunopositive PRL cells (P < 0.0001), but GnRH treatment did not change the proportion of PRL-stained area and even reduced the number of PRL cells (P = 0.0015).
|
, a considerable number of cells were immunopositive for both PRL and GSU after treatment with E2. In contrast, GnRH treatment did not generate any cells double stained for the latter hormones.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
If a cell population in an endocrine gland has a biological function, it is believed that it is under regulatory control by substances from either a distant (hormonal) or a neighboring (paracrine, autocrine) source. Consequently, if a cell population that expresses a defined combination of hormone mRNAs expands on chronic exposure of a regulatory factor and cell populations with other hormone mRNA combinations do not or decreases in size, it is believed that the expanding population has a functional significance. In the present investigation, it was found that chronic treatment with either GnRH or E2 during culture caused significant changes in the number of cells coexpressing particular hormone mRNAs, most notably PRL and GSU mRNA. Both treatments markedly augmented the proportion of cells expressing PRL mRNA together with other hormone mRNAs, but the increase in cells expressing PRL mRNA alone was higher after E2 treatment than after GnRH treatment. Furthermore, GnRH markedly increased both the proportions of cells containing GSU mRNA in combination with other hormone mRNAs and GSU mRNA monohormonal cells, but E2 enhanced only the occurrence of cells showing combinatorial expression of GSU mRNA and not the occurrence of cells expressing GSU mRNA alone. Although culturing pituitary cells resulted in a dramatic fall in the number of both mono- and plurihormonal GH cells, E2 decreased the proportion of monohormonal cells expressing GH mRNA even further but had no effect on the number of plurihormonal GH mRNA cells. All these findings support our proposal that the occurrence of cells displaying combinatorial hormone mRNA expression is specifically regulated, i.e. in a way different from that of monohormonal cells. Because regulation of cell population size is not meaningful if there is no specific function for that population, combinatorial expression of hormone mRNAs appears to serve a specific function and does not seem to represent a random low-expression phenomenon of nonfunctional mRNAs.
Of particular interest was the occurrence of cells showing combinatorial expression of PRL and GSU mRNA without the ßGSU mRNAs or GH mRNA. These cells were very scarce in newly dispersed cells of 14-d-old rat pituitary (<1% of the cells), but their occurrence strongly augmented after treatment with GnRH or E2 in culture, i.e. to about 20% of the total population, strongly supporting the hypothesis that these cells may play a specific role under particular hormonal conditions.
A major aspect of regulated combinatorial expression is whether changes in the expression at the mRNA level match changes at the hormone production level. The present experiments suggest that this is not always the case. GnRH augmented the number of plurihormonal and monohormonal cells expressing PRL mRNA but not the number of PRL-producing cells estimated by immunostaining. Furthermore, this peptide augmented the number of cells coexpressing PRL and GSU mRNA but not the number of cells that double stained for these hormones; in fact, no double-stained cells were found. In contrast, E2 augmented the number of cells coexpressing GSU and PRL mRNA as well as the number of cells coproducing these two hormones. The differential response of plurihormonal cells to GnRH and E2 at the mRNA level vs. that at the protein level indicates that under particular hormonal conditions, combinatorial expression of pituitary hormone genes is primarily at the mRNA level. These findings are consistent with previous findings showing that in 14-d-old rats cells responsive to 3-MSH express hormone genes at the mRNA level in a considerably higher number than cells in which the corresponding hormone was detectable (15).
It should be noted that in the present experimental approach, a chronic treatment regimen was adopted. Because GnRH was used in doses of 1 nM, desensitization in the production of gonadotropin ßGSUs most probably occurred (34, 35). This may have created a nonphysiological condition that resulted in cells expressing GSU without any of the ßGSUs (36). However, even if this were the case, it would not explain the coexpression of PRL with GSU after treatment with GnRH. It also should be noted that the dramatic increase of cells coexpressing GSU and PRL mRNA may be a pharmacological response to the supraphysiological doses of GnRH and E2 used. Nevertheless, the existence of coexpression of these two genes has also been reported in vivo. They were observed sporadically in a previous study in fetal mouse pituitary (16), and chronic treatment with E2 is well known to induce lactotroph hyperplasia (20, 21). That such hyperplasia also seems to occur in our aggregate cell culture system and this is associated with coexpression of GSU in the PRL cells is of potential interest for better understanding the pathogenesis of lactotroph hyperplasia and tumor formation. Indeed, GSU may have an autocrine role in these PRL cells because it stimulates the development (25, 37) and secretory activity of lactotrophs (38). Coexpression of PRL and GSU, but not ßGSUs, has been found in certain prolactinomas that also express the GnRH receptor in the same cells (39).
The increase in proportion of PRL mRNA-containing cells in cultures treated with GnRH may rely on paracrine interactions with gonadotrophs as previously described (29, 30, 37, 40, 41). Alternatively, GnRH may have a direct effect on PRL cells. GnRH receptors have been identified on gonadotrophs but may also be present on lactotrophs because there is uptake of GnRH in these cells by endocytosis (42) and part of the lactotrophs respond to GnRH in terms of intracellular-free calcium rises (Roudbaraki, M., and C. Denef, unpublished observations) (43, 44).
Whether the changes in plurihormonal expression at the mRNA level are due to a proliferative effect on the cells already expressing this particular combination of hormone mRNAs or to de novo expression of one of these hormone mRNAs in cells already expressing the other remains unknown. Based on GnRH immunoneutralization studies, GnRH appears to be involved in the expansion of the gonadotrope cell population, such as after gonadectomy (17, 18, 19) and, of the mitotic cells found in the pituitary after gonadectomy, 80% do not stain for LH or any other pituitary hormone (18), suggesting that cell proliferation may indeed in part be involved and that these proliferating cells may express hormone mRNA but not the hormone protein. That E2 expands the PRL cell population is consistent with literature data showing that this hormone has mitogenic effects on PRL cells (45) and enhances PRL gene expression (46, 47, 48). It is likely, however, that posttranscriptional regulation is also involved (49). Indeed, at least one example has been reported for such mechanism. It has been reported that part of the GH-only-secreting cells contain PRL mRNA but that they can rapidly convert to GH- and PRL-secreting cells under the influence of estrogens in vitro (50), and this was shown to be through a mechanism activating PRL mRNA translation.
The colocalization of PRL mRNA and GSU mRNA in the same cells is important to consider in light of the present concepts of different lineages in the pituitary (51, 52, 53, 54). GSU is expressed in thyrotrophs and gonadotrophs. Most of the thyrotrophs belong, together with the GH and PRL cells, to the Pit-1 lineage. Finding of coexpression of PRL and GSU without ß-subunits or GH is therefore an indication that specification of the different lineages in the anterior pituitary may be less stringent than currently proposed or may be lost in a certain proportion of cells during postnatal life. It has been postulated by Behringer et al. (55) that the majority of PRL cells differentiate from GH-expressing cells, although a small proportion appears to arise from progenitors that do not express GH. Perhaps some GSU cells are these progenitors. GSU mRNA is the first hormonal phenotype expressed in the embryonic rat pituitary (9, 56). It has recently been shown that treatment of fetal mouse with diethylstilbestrol advanced the first appearance of lactotrophs as early as E14, which is earlier than the first detection of GH (57). The coexpression of PRL and GSU mRNA and its up-regulation by GnRH is interesting in view of the report that part of human prolactinomas also coexpress these two genes in the same cells and that only these tumors were responsive to GnRH in terms of PRL secretion (39).
Taken together, the present findings support the hypothesis that cells coexpressing different pituitary hormone genes at the mRNA level are a distinct cell population that do not need extrapituitary factors to survive and can expand or regress by specific regulatory factors. Exposure to E2 or GnRH during culture generates cells displaying combinatorial expression of PRL and GSU mRNA, but only E2 appears to stimulate coproduction of the two hormones. What the precise function of these cells is in vivo remains to be studied, but the present findings at least give an impetus to start such studies.
|
Acknowledgments |
|---|
GSU. Koen Van Acker and Erik Nys are greatly acknowledged for expert assistance with the FACS. We thank Steffen Fieuws from the Biostatistical Centre of the University Medical School (University of Leuven) for the execution of the overall statistical calculations. |
Footnotes |
|---|
Abbreviations: DNase, Deoxyribonuclease; dNTPs, deoxynucleotide triphosphates; E2, estradiol; FACS, fluorescence-activated cell sorter; FITC, fluorescein isothiocyanate; GSU, glycoprotein subunit; LBamp, Luria Broth medium agar containing 100 mg/liter ampicillin; POMC, proopiomelanocortin; PRL, prolactin; RT, reverse transcription.
Received June 11, 2002.
Accepted for publication September 25, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
3-melanocyte-stimulating hormone detected through intracellular Ca2+ responses in immature rat pituitary constitute a fraction of all main pituitary cell types, but mostly express multiple hormone phenotypes at the messenger ribonucleic acid level. Refractoriness to melanocortin-3 receptor blockade in the lacto-somatotroph lineage. Endocrinology 140:4874–4885-subunit and LHß messenger ribonucleic acid levels in rat anterior pituitary cells in culture. Mol Endocrinol 2:521–527[Abstract]
subunit. Science 226:566–568-subunit and daily gonadotropin releasing hormone treatment on prolactin release from 50-day ovine foetal pituitary explants. J Neuroendocrinol 13:199–208[CrossRef][Medline]
in the anterior pituitary gland. Mol Endocrinol 11:674–681
