Ghrelin Restoration of Function In Vitro in Somatotropes from Male Mice Lacking the Janus Kinase (JAK)-Binding Site of the Leptin Receptor

Abstract

Deletion of the signaling domain of leptin receptors selectively in somatotropes, with Cre-loxP technology, reduced the percentage of immunolabeled GH cells and serum GH. We hypothesized that the deficit occurred when leptin's postnatal surge failed to stimulate an expansion in the cell population. To learn more about the deficiency in GH cells, we tested their expression of GHRH receptors and GH mRNA and the restorative potential of secretagogue stimulation in vitro. In freshly plated dissociated pituitary cells from control male mice, GHRH alone (0.3 nM) increased the percentage of immunolabeled GH cells from 27 ± 0.05% (vehicle) to 42 ± 1.8% (P < .002) and the secretion of GH 1.8–3×. Deletion mutant pituitary cells showed a 40% reduction in percentages of immunolabeled GH cells (16.7 ± 0.4%), which correlated with a 47% reduction in basal GH levels (50 ng/mL control; 26.7 ng/mL mutants P = .01). A 50% reduction in the percentage of mutant cells expressing GHRH receptors (to 12%) correlated with no or reduced responses to GHRH. Ghrelin alone (10 nM) stimulated more GH cells in mutants (from 16.7–23%). When added with 1–3 nM GHRH, ghrelin restored GH cell percentages and GH secretion to levels similar to those of stimulated controls. Counts of somatotropes labeled for GH mRNA confirmed normal percentages of somatotropes in the population. These discoveries suggest that leptin may optimize somatotrope function by facilitating expression of membrane GHRH receptors and the production or maintenance of GH stores.

Somatotropes are regulated throughout life by an ensemble of hormones, including hypothalamic GHRH, ghrelin, and somatostatin, sex steroids, T₄, and glucocorticoids (1–3). They also express receptors for the anorexic peptide, leptin (4). These leptin receptors (LEPRs) are related to the class 1 cytokine receptor superfamily (5, 6). The long form of the receptor, LEPRb, has a single-pass transmembrane domain and a 302-amino acid cytosolic domain that binds and activates the Janus kinase (JAK)/signal transducer and activator of transcription pathway (7). After autophosphorylation of JAK2, tyrosine 985 is phosphorylated, which activates the phosphoinositol-3-kinase pathway. JAK2 phosphorylation stimulates the phosphorylation of Tyr1138 at a site that attracts the binding and activation of signal transducer and activator of transcription 3, which activates gene transcription. Despite the fact that more than 80% of somatotropes express LEPRb (4), the importance of leptin and these JAK-mediated pathways to somatotrope function is not well understood.

Most of the early in vivo evidence for a stimulatory effect of leptin on somatotropes comes from studies of ob/ob or db/db mice, in which the numbers of somatotropes and serum GH are reduced (8, 9). However, the cause of this reduction in GH expression is not known. It could be due to the lack of direct leptin stimulation. Alternatively, one may also argue indirect causes from the fat load, metabolic disease, and/or the hypogonadal condition of these mice, because steroid stimulation of GH cells is needed for their maintenance (7, 10, 11). Recently, Luque et al (12) reported that leptin restored GH secretion in ob/ob mice, which were infused with exogenous leptin for 7 days. In addition, leptin caused an increase in GHRH receptors. No changes were noted in GHRH levels themselves, which suggested that GHRH did not mediate the restoration.

Studies of leptin's effect on somatotropes agree that the LEPRs are functional, but they do not agree about the direction of its effect. The reports show that leptin may stimulate or inhibit GH, depending on experimental conditions and species (8, 12–24). In vitro studies from our laboratory reported that, in rats, exposure to 10–100 pg/mL leptin restored the population of somatotropes reduced by 24 hours of fasting (25), and 1–10 pg/mL of leptin stimulated an increase in GH mRNA in normal pituitary somatotropes. This suggested that, in relatively low doses, leptin may be important to the maintenance of the pituitary somatotrope population.

To further investigate the direct role of leptin in optimizing somatotrope functions, we used Cre-LoxP technology to selectively delete exon 17 of the LEPR gene in somatotropes. Exon 17 codes for the JAK binding site (4). Because all signaling pathways are activated by JAK2 binding (26–29), this removes all known possibilities for leptin signaling to the target cells. The deletion mutant mice bearing GH cells that could not receive leptin signals showed no LEPRb proteins and, in fact, LEPRb proteins were reduced overall in the pituitary (4). The mice were GH deficient and expressed increased adiposity by 5–6 months of age (4). At the level of the pituitary, there was a significant, 60%, reduction in numbers of GH cells, which correlated well with the lower serum GH. Thus, our first set of in vivo studies suggested that leptin may be needed to either increase the population of somatotropes during postnatal development (30, 31), or possibly to optimize and maintain the population in the adult, or both.

However, before we could test this working hypothesis, we recognized that further work was needed to characterize the deficit itself. Were the somatotropes really “missing”? If not, we predicted that they could be detected by other cytochemical means. Furthermore, we postulated that their GH stores and secretion might be restored by other somatotrope-stimulatory hormones. The fact that we did not detect reduced overall numbers of cells or a reduction in size in the anterior pituitary of mutants (4) supported the hypothesis that cells with the potential to be somatotropes might be present.

The series of studies in this report focused on testing this hypothesis. The study was designed in two phases over a 2-year period. The first phase determined whether the mutant GH cells were responsive to GHRH and whether they had adequate expression of GHRH receptors. The latter test was done by counting cells labeled by affinity cytochemistry after exposure to biotinylated analogs of GHRH (32). This helped determine whether reduced membrane receptor expression was one possible mechanism behind the reduced expression of GH. After we discovered significant reductions in responses to GHRH and expression of GHRH receptors, we began the second phase of the study during the second year. We also experimented with plating density, to optimize responses of the cultures to secretagogues.

During the second phase, we tested the additive effect of ghrelin on GHRH-stimulated secretion. The rationale for this was not based on a known ghrelin defect, because we had previously reported that circulating ghrelin levels were normal in the male mutant mice (33). Furthermore, whereas ghrelin does enhance GH secretion (34–36), circulating levels are not believed to drive GH pulses normally (37). Ghrelin was used mainly to help detect quiescent somatotropes, by testing for the presence of cells with GH secretagogue receptors (GHS-R) and activating signaling pathways not activated by GHRH (34, 35), which might restore the cells' functions. Once we discovered that ghrelin could restore GH cell functions, we confirmed the presence of quiescent GH cells with a newly developed in situ hybridization protocol that detected GH mRNA in the mouse. This report thus presents a number of new discoveries about the cellular basis behind the deficiency in the mutant somatotropes, which are critical to the design of future mechanistic studies of the leptin's optimization of GH cell functions.

Materials and Methods

Animals

Deletion mutant mice with LEPRs specifically deleted from somatotropes were produced as previously described (4). Deletion mutant male mice and littermate controls (3 to 5 months old) were taken before development of obesity to prevent confounding affects of the fat load. There was no difference in weight when 15 mutants (31.4 ± 0.5 g) and 15 littermate controls (29.2 ± 1) were compared.

They had been housed four to five animals/cage with a 10-hour light, 14-hour dark cycle in an environment maintained at a constant room temperature of 68°F. Mice were fed Teklad 8640 rodent diet (Teklad, Madison, Wisconsin), which contained 22% protein, 5% fat, and 4.5% fiber. The animal care protocol (nos. 3014 and 3337) was approved annually by the institutional animal care and use committee.

Collection of pituitary cells

Isoflurane-anesthetized mice were decapitated, and pituitaries from mutant and littermate control mice were quickly removed and dispersed into single cell suspension with the recently optimized protocol described previously (4). In all studies (phase I and II), cultures from mutant and littermate control mice were collected before 10:00 am. After overnight culture, cells from the two experimental groups were stimulated at the same time. In the first phase of the study (four replicated experiments), freshly dispersed pituitary cells were plated at 20 000 cells per well on poly-d-lysine-coated coverslips in 24-well tissue culture plates and were allowed to grow overnight in DMEM inside a 5% CO₂ incubator set at 37°C. During the second phase of the study (four replicated experiments), we determined that a lower plating density (12 000 cells per well) actually facilitated a more sensitive response to GHRH in both control and mutant cultures.

Cells were exposed to either GHRH (0–30 nM), leptin (20 nM), or the combination of (0–10 nM) GHRH and 10 nM ghrelin for 3 hours. Lower concentrations of ghrelin alone were not tested, because the cells were only responsive to 10 nM levels. After stimulation, media from both mutant and littermate control cultures were collected for hormone (GH) immunoassay, and the cells were fixed in 2% glutaraldehyde for 30 minutes at room temperature followed by four washes with 0.1 M phosphate buffer containing 4.5% sucrose and 0.15% glycine. Fixed cells were kept at 4°C for immunolabeling or in situ hybridization experiments. Immunoassays were run the same day on fresh media if at all possible.

Single labeling for GH

Monolayers of glutaraldehyde-fixed pituitary cells from mutant and littermate control cultures were immunolabeled at the same time for GH with 1:225 000 anti-rGH (A. Parlow, National Institute of Diabetes and Digestive and Kidney Diseases Hormone Distribution Office) with the use of a protocol described earlier (7).

Dual labeling with biotinylated GHRH and GH

Pituitary cells from both mutant and littermate control cultures plated on glass coverslips were treated with biotinylated GHRH for 10 minutes at 37°C before fixation with glutaraldehyde for 30 min. Fixed cells were then labeled with avidin-biotin peroxidase complex (ABC kit, Vector laboratories, Inc, Burlingame, California), and the peroxidase was detected by a black diaminobenzidine substrate with the use of a protocol described earlier, followed by immunolabeling for GH (32). In all cytochemistry experiments, both mutant and littermate control cells were stimulated and labeled at the same time, with the same set of solutions.

In situ hybridization

Glutaraldehyde-fixed cells from mutant mice and littermate controls were used for in situ hybridization to detect GH mRNA using a modification of the previously published protocol (38, 39), which had been used with rat pituitary cells. The same probes were used in both protocols (rat and mouse). Antisense and sense oligonucleotide probe for mouse GH mRNA were made by GeneDetect.com. The antisense probe hybridizes to nucleotides 589–636 of the mouse GH gene mRNA (GenBank accession no. NM_008117). The sequence for the antisense probe is TTCTTGAAGCAGGAGAGCAGCCCATAGTTTTTGAGCAGCGCGTCGTCG. Cells from mutant or littermate control cultures were incubated in GH probe overnight (17–18 h) at 35.6°C. Optimal labeling was obtained with 3 ng/mL of GH probe. The same amount of sense GH probe produced no labeling.

Cells containing the biotinylated probe for GH mRNA were detected by immunocytochemistry for biotin, with a modification of a previously published protocol, which had been originally developed for rat tissues (11, 38, 39). The modification was needed because of the presence of monoclonal antibiotin (made in a mouse) in the original protocol designed for rat pituitaries. The changes are as follows. We substituted goat antibiotin for the monoclonal (mouse) antibiotin in the first detection step. This change necessitated the substitution of biotinylated horse antigoat IgG for the biotinylated horse antimouse IgG used in the earlier protocol. The dilution of both solutions remained 1:100.

After the probe was applied overnight to both mutant and littermate control cells, the plated cells were washed as in the previous protocol, and the biotin was then detected by the following steps: step 1 involved the application of a blocking solution containing 10% normal horse serum for 15 minutes; step 2 applied 1:100 dilution of goat antibiotin (Vector Laboratories) for 30 minutes at 37°C in a shaking incubator. After washing in Tris-buffered saline, step 3 involved the application of 1:100 biotinylated horse antigoat serum (Vector Laboratories) for 10 minutes followed by a TBS wash. Steps 2 and 3 were then repeated. Finally, the biotin in the complex was detected by 1:100 streptavidin peroxidase for 10 minutes. The peroxidase was detected by nickel-intensified diaminobenzidine, as in our previous studies (11, 38, 39). This gave a punctate, dense black reaction product in somatotropes. After the in situ hybridization reaction, a subset of the cells was immunolabeled for GH as in previous studies (11, 38, 39). The second labeling protocol produced an orange/amber reaction product. This dual labeling protocol has been tested in method and specificity controls in previous studies.

GH immunoassay

All collected media samples were assayed the same day they were collected or, if that wasn't possible, they were frozen until the assay. More frequently, they were immediately assayed for GH using Luminex LX200 (Luminex Corp, Austin, Texas) xPONENT 3.1 with Milliplex MAP Mouse Singleplex GH kit (Millipore Corp, Billerica, Massachusetts) following the manufacturer's instructions. Media samples from both control and mutant cultures were diluted 1:30 in DMEM and the dilution buffer supplied with the kit. Each sample was assayed in duplicate wells, and all assays contained samples from both mutant and control wells.

Statistics

Power analysis

Each experiment involved at least three animals, and a given result will be the average of four replicates of that experiment. To establish the number of replicates needed for each experiment, a post hoc power analysis was done with pilot studies. This allows us to predict that a 47% reduction in GH from 50 ng/mL (Controls) to 26.7 ng/mL (in the deletion mutants) with an SD of 11 ng/mL will have 95.6% statistical power with three replicates in a two-tailed, 0.05 level t test. http://www.dssresearch.com/KnowledgeCenter/toolkitcalculators/statisticalpowercalculators.aspx.

Statistical analysis

Cell counts and assay values were analyzed using InStat3 or PRISM 5.0 software. The Kruskal-Wallis, one-way ANOVA was run to detect significant differences between mutant and control mice, or between experimental groups in the secretion studies. This was followed by Tukey's and Bonferroni's multiple comparison post hoc tests to determine which changes were significant (P < .05 was considered significant). When two experimental groups were compared, the Mann-Whitney t test was run.

Results

Mutant cultures secrete less GH and are poorly responsive to GHRH

The first set of four replicate experiments was designed to determine whether stimulation with GHRH for 3 hours would increase the GH stores, so that we would detect an increase in the percentages of immunolabeled GH cells in cultures from deletion mutants. Figure 1 shows the averaged counts (Figure 1A) and the photographs of immunolabeled cells (Figure 1, B–G). Control cells had 28% cells immunolabeled for GH whereas, in agreement with our previous studies (4), mutant cultures had only 16% immunolabeled GH cells (which is significantly different). Stimulation with 3 nM GHRH or 20 nM leptin caused a significant increase in percentages of GH cells in control cultures to 42% or 35%, respectively. The same concentrations had no effect on percentages of mutant cells. The top panel in the photomicrographs (Figure 1, B–D) illustrates the increase in immunolabeled GH cells with GHRH stimulation in the control cultures, and the bottom panel illustrates the lack of change in the cultures from mutant animals (Figure 1, E–G).

Figure 1.

Counts of Immunolabeled GH Cells after GHRH Stimulation. A, Treatment of control cultures with 0–30 nM GHRH or 20 nM leptin stimulates an increase in the percentages of cells with GH proteins. No changes are seen in cultures from mutant animals. Star, significantly different from vehicle control values at P = .028 (0.3, 30 nM GHRH); P = .037 (3 nM GHRH); P = .024 (10 nM GHRH); P = .04 (20 nM leptin). Double star, significantly different from control culture values (same treatment group) (P < .03). B–D, Immunolabeled fields (for GH) showing changes in control cultures in the presence of vehicle (panel B), 3 nM GHRH (panel C), or 20 nM leptin (panel D). The labeled cells have dark gray-black label. The mutant cell cultures are depicted in panels E–G. Panel E shows a few small cells in the vehicle control. The mutant cultures do not show an increase in immunolabeled GH cells in any of the concentrations of GHRH (3 nM is depicted in panel F), or in the presence of leptin (panel G). Arrows, labeled cells; bar, 20 μm.

Figure 2 illustrates the changes in media GH from this first set of experiments. Despite the increase in GH cells (Figure 1), control cultures did not secrete more GH in the presence of 0.3 nM GHRH (data not shown). However, there was a steady dose-related increase in GH in 3–30 nM of GHRH from 1.8 to 3× control values. Neither mutant or control cultures secreted more GH in response to 20 nM leptin. The mutant cultures secreted less GH basally and in the presence of all concentrations of GHRH. Only 30 nM GHRH stimulated a significant 2.3-fold increase in GH from mutant cultures. Because of the low starting secretory levels (basal), the stimulation only resulted in levels that were similar to basal levels secreted from control cultures.

Figure 2.

Effect of GHRH on the Secretion of GH from Control or Deletion Mutant Pituitary Cells. Control cells respond to 3–30 nM GHRH (star, P = .048, 3 nM; P < .0027, 10 nM; P = .001, 30 nM). Deletion mutant cells secrete lower basal levels of GH when compared with control cultures (double star, P = .01, 0 nM; P = .002, 3 nM; P = .0035, 10 nM; P < .001, 30 nM). However, in mutant cultures, 30 nM GHRH stimulates higher levels of GH when compared with vehicle (star, P < .0001). Leptin does not stimulate GH secretion from either set of cultures (data not shown).

Mutant cultures have fewer cells expressing GHRH receptors

We also determined that the lack of responsiveness to GHRH correlated with a reduced expression of GHRH receptors. We applied cytochemical studies to detect available surface receptors for GHRH. Control and deletion mutant cells were exposed to biotinylated analogs of GHRH for 10 minutes while living, which has been shown, in previous studies, to bind the GH population, maximally, with sufficient surface labeling for the full detection of the target population (32). After fixation, the biotinylated analog is detected with avidin biotin peroxidase complex (Vector Laboratories). The reaction product for peroxidase (nickel-intensified diaminobenzidine) is black in patches over the cell and can easily be distinguished from the orange immunolabeling for GH (32). Examples of labeling for biotinylated GHRH are shown in Figure 3. The dense black label represents the biotinylated GHRH, and the gray cytoplasmic labeling (orange photographs) marks sites of GH proteins. Figure 3D is a graph showing the results of counts of the cells expressing biotinylated GHRH. Normally, the percentages of GHRH target cells are similar to those of the GH cell population (25%–30%) and control cultures show this clearly. In contrast, the mutant cultures show a significant reduction in GHRH target cells to 12.3% of pituitary cells.

Figure 3.

Dual Labeling for Biotinylated GHRH (black, arrows) and GH Proteins (orange or gray in the black and white image) in Control (panels A and B) and Deletion Mutant (panel C) Cultures after Exposure to 1 nM Biotinylated GHRH for 10 Minutes. D, Graph of the counts of percentages of cells labeled for biotinylated GHRH, showing a significant reduction in mutant cultures. Star, significantly different from control, P = .0003.

Ghrelin added to GHRH will restore numbers of immunolabeled GH cells in mutants to normal levels

Phase 2 studies were designed to determine whether ghrelin, another secretagogue for somatotropes, would restore or rescue the mutant GH cell population alone or in combination with GHRH. Ghrelin was essentially used as a test to determine whether cells bearing GHS-Rs were still present in the population. During this phase, we plated the control and mutant cells less densely (12 000 cells per well), which produced better secretory responses to lower concentrations of GHRH.

Figure 4A shows the results of counts of immunolabeled GH cells in control or mutant cultures after 3 hours in 10 nM ghrelin alone, or after treatment with 10 nM ghrelin added to 0–10 nM GHRH. In control cultures (first bar graph), GHRH stimulated increases in immunolabeled GH cells, as expected. In the second bar graph, the first data point shows that 10 nM ghrelin alone did not change the percentages of immunolabeled GH cells. The remaining points on the bar graph show the effects of combining 10 nM ghrelin with 0.1–10 nM GHRH. Ghrelin stimulated more GH cells when it was added with 0.1 and 1 nM GHRH.

Figure 4.

Counts of Immunolabeled GH Cells after GHRH and Ghrelin Stimulation. A, Graph shows changes in the percentage of immunolabeled GH cells in response to GHRH with and without 10 nM ghrelin. Both control and mutant cultures were exposed to different concentrations of GHRH (0–10 nM). Half of these cultures also received 10 nM ghrelin. Cultures were plated less densely and, as a consequence, they showed enhanced secretory and storage responses, possibly due to the reduced ultrashort loop feedback by GH. The first bar graph (control cells) shows that, with the Mann-Whitney test, all concentrations of GHRH stimulated a significant increase in GH cells over vehicle (filled star, P = .05, 0.1 nM GHRH and 1 nM GHRH; P < .04, 0.3 nM and 10 nM GHRH). The first data point in the second bar graph (control + ghrelin) showed that 10 nM ghrelin alone did not enhance GH secretion. However, ghrelin did enhance the stimulatory effect of 1 and 0.1 nM GHRH (open stars, P < .04). The third bar graph (mutant cells) showed the significantly reduced percentages of GH cells in mutant cultures; the values are lower than all other values (one-way ANOVA, open circles), and neither 0.1 nor 0.3 nM GHRH alone stimulates an increase in percentages of GH cells. Adding 1–10 nM GHRH alone to mutant cultures stimulated a slight increase in percentages of GH cells over basal levels (filled circles, P < .038). The first data point in the fourth bar graph showed that the percentages of GH cells were increased in populations from deletion mutants in the presence of 10 nM ghrelin alone (filled circle, P = .01). The remaining data points in the fourth bar graph showed that cultures exposed to 10 nM ghrelin with 0.1–10 nM GHRH had striking increases in the percentages of GH cells (closed triangles, P < .05, 0.1, 10 nM GHRH; P < .038, 0.3, 3 nM GHRH). Filled star, significantly different from vehicle control; open star, significantly different from cultures treated with same dose of GHRH only; open circles, significantly different from all other values (the lowest values); filled circles, significantly different from mutant cultures receiving vehicle only; triangle, significantly different from culture treated with same dose of GHRH only. Levels secreted in the presence of ghrelin and 0.1–10 nM GHRH were indistinguishable, statistically from levels seen in control wells exposed to the same concentrations of GHRH. B and C, Immunolabeling for GH comparing fields from control (panel B) and deletion mutant (panel C) cultures after treatment with 3 nM GHRH and 10 nM ghrelin. The increase in the mutant cultures has rendered the population of immunolabeled GH cells indistinguishable from control cultures. Bar, 20 μm; arrows, immunolabeled cells.

The third bar graph in Figure 4A shows that the mutants responded better to GHRH alone. When they were plated 12 000 cells per well, there was a slight, but significant, increase in percentages of immunolabeled GH cells (over mutant cells exposed to vehicle only) after exposure to 1–10 nM GHRH alone. The fourth bar graph in Figure 4A shows that 10 nM ghrelin alone also stimulated a slight, but significant, increase in percentage of immunolabeled GH cells in mutant cultures. However, the most striking increase leading to recovery or restoration of immunolabeled GH cells was seen when 10 nM ghrelin was added with GHRH (the remaining points in the fourth bar graph). All changes in the fourth bar graph were significantly different from the mutant cultures that received the same concentration of GHRH only. In the presence of 10 nM ghrelin and 3 or 10 nM GHRH, the GH cell percentages increased to levels that were not different from those seen in control cultures treated with GHRH only. This indicated a complete recovery of GH stores in the population. Figure 4, B and C, shows representative fields from control (Figure 4B) and mutant (Figure 4C) cultures treated with 3 nM GHRH and ghrelin showing that, in terms of numbers of immunolabeled somatotropes, the mutant cultures were now indistinguishable from control cultures.

GH mRNA-bearing cells are present in mutants; most are “invisible” to immunolabeling

The restoration of immunolabeled GH cells in the population by the combined treatment with ghrelin and GHRH suggested that GH cells with the potential to produce GH proteins were present. We had hypothesized that stores of GH and expression of GHRH receptors in mutants may have dropped below threshold levels needed for their detection. We further hypothesized that the cells might be detected by their content of other markers for somatotropes, the most important being GH mRNA. After modifying the in situ hybridization technique for use with mouse tissues, we applied it to freshly plated, fixed cells from control and mutant mice.

Figure 5 shows the counts of cells with GH mRNA, demonstrating only a slight reduction in numbers of somatotropes in the mutant cultures. Figure 5, B and C, illustrates control and mutant cultures, respectively, containing similar levels of GH mRNA-bearing cells. Figure 6 depicts the actual deficiency in GH protein storage in dual-labeled fields comparing normal GH stores (orange label) in control (panels A and B) mice with the low or absence of GH stores in the mutant (panels C and D) mice. In essence, control mice GH cells have the typical black (mRNA) and amber/brown (GH antigens) labeling pattern. In contrast, the mutant GH cells have strong labeling for the mRNA but very little of the label for GH antigens.

Figure 5.

Counts and Illustrations of Cells Bearing GH mRNA. A, Control cultures labeled with in situ hybridization techniques for GH mRNA; GH cells are shown by patches of dense label. B, Mutant culture with a labeling pattern for GH mRNA similar to that of the controls. C, Graphs show counts of freshly dispersed control and mutant cultures showing similar percentages of cells labeled for GH mRNA. The slight reduction (star, P = .0003) is significant; however, the percentages are still within the previously reported range for GH cells (7, 39). Bar, 20 μm; arrows, labeled cells.

Figure 6.

Dual Labeling for GH mRNA by in Situ Hybridization (black, arrows) and GH Proteins by Immunolabeling (orange-brown). A–C, Fields showing labeling in control mouse pituitary cells. D and E, Labeling in mutant cells. All of the GH cells in the control cultures have both GH mRNA (black) and proteins (orange-brown). In contrast, most of the mutant cells show GH mRNA (black) but little GH proteins (orange-brown). Arrows point to cells with patches of GH mRNA in the cytoplasm. Bar, 20 μm.

Ghrelin also restores secretory responses of mutant cells to GHRH

When media from four different experiments testing the additive effects of ghrelin were analyzed, none of the control cultures showed a significant additive effect of ghrelin on GH secretion (data not shown). Therefore, Figure 7 only shows the dose-dependent responses to GHRH alone (first bar graph). The second bar graph showed that mutant cultures secreted less GH than control cultures, and only 1 and 3 nM GHRH effectively stimulated secretion slightly, with levels significantly lower than those of stimulated control cultures. In the third bar graph, the first two data points show that 10 nM ghrelin did not stimulate GH secretion by itself, or when added with 0.1 nM GHRH. However, there was a dose-dependent increase in GH secretion in the presence of 10 nM ghrelin added with 0.3–3 nM GHRH. Ghrelin and 1 or 3 nM GHRH completely restored the GH secretion in the population, because the resulting GH levels were not different from those of stimulated control cultures. This correlates well with data in Figure 4.

Figure 7.

GH Secretion from Control and Mutant Cells Plated Less Densely Than Those in Figure 1. The first bar graph shows the response from control cells to increasing concentrations of GHRH. There is a significant increase after 0.3–1 nM GHRH (filled star, P < .0001) and 10 nM GHRH (P = .0002). Ghrelin (10 nM) did not augment GH secretion from control cells either alone or with GHRH. These data are not included to simplify the graph. The second bar graph showed that secretion from mutant cultures was still lower than all other values (star, P = .0009). However, mutant cultures did respond slightly to 1 nM (filled triangles, P = .0002) and 3 nM (P < .0001) GHRH alone. In the third bar graph, when values from each dose of GHRH were compared, mutant cells exposed to GHRH alone secreted less GH than control cells (filled circles, 0.1 nM, P = .0004; 0.3 nM, P < .0001; 1 nM, P = .035; 3 nM, P = .026). In the fourth bar graph, ghrelin alone neither stimulated GH secretion from mutant cultures nor did it augment the effect of 0.1 nM GHRH. However, ghrelin did augment the stimulatory effect of 0.3 nM (P < .0001), 1 nM (P = .0003), and 3 nM (P < .0001) GHRH (open triangles). Mutant cultures stimulated with ghrelin and 1 or 3 nM GHRH appeared completely recovered because levels of GH were not different from those seen in control cultures stimulated with the same concentrations.

Discussion

The overall objective of these ongoing studies of this mouse model (4, 33) has been to determine the significance of leptin regulation of somatotrope functions. The first set of studies characterized the phenotype of the somatotrope LEPR^exon17 null mice, which included GH deficiency and adult-onset obesity (4). Follow-up metabolic studies of preobese mutant mice showed that the obesity stemmed from reductions in fat burning and activity, and changes in adipokines that were the result of the GH deficiency (33). Thus, we showed that leptin does play a significant role in the optimization of somatotrope functions with respect to normal body composition.

Once we established the significance of leptin regulation of somatotropes, many questions about the mechanism and timing of leptin's actions on somatotropes remained. We had originally hypothesized that leptin was important to somatotropes as early as postnatal development (30, 31) and that the postnatal surge of leptin was responsible for the expansion in the somatotrope population during postnatal life. This hypothesis assumed that the deficiency in the somatotrope population seen in the adult (4, 33) reflected a loss, resulting in fewer cells to support GH functions. However, this assumption had never been proved. Furthermore, total dispersed pituitary cell counts and the size of the anterior pituitary had already suggested that cells with somatotrope potential might not be missing (4).

Thus, before we could design a series of studies to determine mechanisms behind leptin actions, it was clear that we needed to know more about the deficiency in the somatotropes themselves. The purpose of the present report was to test the hypothesis that the GH cells were present, albeit quiescent. If this proved correct, then we hypothesized that they could be detected either by their expression of another somatotrope product (GHRH receptors or GH mRNA) or they might be revealed by the activation of another set of receptors, such as the GHS-R.

The present study was designed in two phases to test these hypotheses. The overall objective of the first phase of this study was to determine whether GHRH stimulation alone would rescue somatotropes, perhaps by restoring GH hormone stores needed for secretion. Even when we allowed 3 hours for stores to build, the results showed that the mutant somatotropes responded poorly to GHRH both in terms of secretion and storage of GH. This discovery was then correlated with the detection of an overall reduction in the expression of GHRH receptors in the population. Both of these discoveries correlated well with studies by Luque et al. (12), who rescued GH functions in ob/ob mice by treating them for 7 days with leptin. They were able to recover both secretion of GH and GHRH receptor expression. Perhaps leptin plays an important role in the production and/or maintenance of GHRH receptors and the GH secretory potential.

After GHRH by itself was inadequate in restoring the GH cell population, we set up phase 2 studies. We changed the plating density, testing the hypothesis that reduced responses might be the result of ultrashort loop negative feedback on GH secretion after the build up of GH in the media. The reduction in plating to 12 000 cells per well in 24-well trays was successful in that it allowed us to detect secretion with lower concentrations of GHRH in both mutant and control cultures. This facilitated the studies of the additive effect of ghrelin. Ghrelin is another secretagogue for somatotropes, which activates via the GHS-R (36, 40, 41) and different signaling pathways (34, 35).

Ghrelin and GHRH together produced an additive stimulatory effect on the percentages of immunolabeled GH cells in control cultures. The effects of ghrelin on percentages of immunolabeled GH cells in mutant cell populations were much more striking, however, probably because of the much lower baseline starting point. Ghrelin alone had a slight effect on the population's GH cells; however when it was added with 1–3 nM GHRH, ghrelin completely restored the population of immunolabeled GH cells to percentages that were not different from those of stimulated control cultures.

These discoveries are significant for two reasons. First, they show that the somatotropes with GHS-R were indeed present in the mutant population and capable of being activated. This refutes our original hypothesis, which stated that the postnatal surge of leptin was needed for the postnatal expansion of somatotropes. In addition, the present study now shows that the somatotrope population in the mutant mice may be quiescent, storing levels of GH proteins and expressing levels of GHRH receptors below threshold levels needed for detection by cytochemistry. Both of these changes correlate with the lower serum GH found in vivo (4, 33).

To confirm the presence of the quiescent somatotropes, we reasoned that other markers, such as GH mRNA, might be used to detect the population and, therefore, we developed modifications of in situ hybridization protocols for the mouse. Our hypothesis was confirmed by the normal numbers of GH mRNA-bearing cells. The slight, but significant, reduced percentages of cells with GH mRNA in the mutants are still well within the range normally reported for GH mRNA-bearing cells (7, 39).

We evaluated the results of assays of media GH to determine whether ghrelin could rescue the mutant population's secretory activity. As reported by others (34), ghrelin did not add to the effect of GHRH in control cultures. However, there were clear additive effects of ghrelin on secretion from mutant cultures stimulated by 0.3–1 nM GHRH, and levels of GH secreted in response to ghrelin and 3 nM GHRH were indistinguishable from those of control cells. Thus, the increased number of GH-storing somatotropes brought out by ghrelin and GHRH (Figure 4) were clearly able to function normally and secrete normal levels of GH.

What do these in vitro discoveries tell us about the normal functions for ghrelin and GH secretion in vivo? We selected ghrelin mainly to test for the presence of the somatotropes in the mutant, because of its broad set of signaling pathways (34, 35). We must recognize that ghrelin cannot be considered a “substitute” for leptin, because it could be acting through a number of pathways that are different from those stimulated by leptin, although one of the pathways for ghrelin involves phosphoinositol-3-kinase, which is also activated by JAK2.

Furthermore, with respect to physiologic relevance, we know that that these experiments are not correcting for an in vivo defect in circulating ghrelin in these mutants. We had previously assayed serum ghrelin in these mutant mice, finding levels to be normal in the males (33). Therefore, circulating ghrelin is not able to compensate for the lack of leptin signaling to mutant somatotropes in vivo. This would be expected, because studies have shown that serum ghrelin does not normally play a role in pulsatile GH release (37). According to these studies, ghrelin levels normally vary throughout the day, peaking just before mealtime, and there is no correlation between GH pulses and ghrelin's peak activity. Nevertheless, studies by Sun et al. (36) of GHSR-null mice, confirmed that ghrelin's receptor was biologically relevant in its stimulation of GH release.

Moreover, it has been suggested that the most relevant source of ghrelin for normal somatotropes is probably that produced by the pituitary somatotropes themselves (42, 43), which suggests that the somatotropes are dependent on its action via an autocrine/paracrine role. Kamegai et al. (42) reported a correlation between hypothalamic GHRH expression and pituitary ghrelin expression (mRNA and protein content). They also showed that the GHS-R inhibitor significantly reduced GHRH-mediated GH release, but not GH release, in the absence of GHRH. Collectively, these data suggested that pituitary ghrelin is involved in optimizing the somatotrope responsiveness to GHRH and that both pituitary ghrelin expression and ghrelin's actions are dependent on GHRH.

These studies thus reveal a major clue about the actions of ghrelin in our mutants. Because the GH cells are GHRH receptor deficient, they may be pituitary ghrelin deficient. The in vitro studies in the present report may thus have restored sufficient levels of ghrelin to optimize responses to GHRH. The mutant cells had what was needed for restoration: both GHS-receptors and GH mRNA. Looking at the immunolabeling evidence in Figure 4, one can see that ghrelin alone stimulated an increase in the number of immunolabeled somatotropes in mutants, bringing GH stores to near-normal control levels. Then, ghrelin and 0.1 nM GHRH together successfully increased stores, to bring the detectable population up to near-normal percentages of immunolabeled GH cells. The population was increased further, in response to higher concentrations of GHRH and ghrelin. Correlating data from Figures 4 and 7 showed that, in the presence of ghrelin, the numbers of immunolabeled GH cells were high enough to support normal GHRH-stimulated levels of GH secretion. These discoveries now set the stage for future studies of mechanisms behind the ghrelin-mediated restoration, including tests of the hypothesis that the mutant cells are ghrelin deficient.

Whereas the major objective of the study was to detect and restore the somatotrope population, the results do provide the basis for future studies of mechanisms behind leptin actions. The loss of LEPRs in the mutant somatotropes appeared to have removed the optimizing influence of leptin on GH cell stores, which is needed to prevent GH deficiency. This could be direct, or indirect through the loss of membrane GHRH receptors. It is tempting to speculate that the mechanism involves a deficiency at the level of regulation of GH translation. However, there are a number of regulatory gateways for trafficking through the rough endoplasmic reticulum or Golgi complex that might also be defective, causing reduced GH storage.

To summarize, these in vitro studies show that ablation of the LEPRs reduced cellular expression of GH proteins and GHRH receptors, with little effect on GH mRNA. The relatively low stores of GH and reduction in receptivity to GHRH led to GH deficiency with metabolic consequences (4, 33). The present study showed that GHRH could not restore GH stores alone, but ghrelin and GHRH together were successful in normalizing both GH storage and secretion. Thus, we discovered that the GH cells were still present in the mutants based on their expression of GH mRNA and GHS-Rs. These new discoveries set the stage for future studies of mechanisms behind pituitary ghrelin's actions as well as potential sites of action for leptin.