- Split View
-
Views
-
Cite
Cite
T. Clark Brelje, Laurence E. Stout, Nicholas V. Bhagroo, Robert L. Sorenson, Distinctive Roles for Prolactin and Growth Hormone in the Activation of Signal Transducer and Activator of Transcription 5 in Pancreatic Islets of Langerhans, Endocrinology, Volume 145, Issue 9, 1 September 2004, Pages 4162–4175, https://doi.org/10.1210/en.2004-0201
- Share Icon Share
Abstract
Although the β-cells of the pancreatic islets of Langerhans express both prolactin (PRL) and GH receptors, we have observed that PRL is considerably more effective than GH in the up-regulation of islet function in vitro. This study examined whether differences in the activation of the Janus kinase 2/signal transducer and activator of transcription (STAT) 5 signaling pathway by these closely related receptors may be involved in this disparity. The activation of STAT5B by PRL was biphasic, with an initial peak within 30 min, a nadir between 1 and 3 h, and prolonged activation after 4 h. In contrast, the response to GH was transient for 1 h. The importance of the long-term activation of STAT5B by PRL was supported by the similar dose response curves for STAT5B activation and the PRL-induced increases in insulin secretion and islet cell proliferation. Because the pulsatile secretion of GH affects its actions in other target tissues, the ability of pretreatment with either hormone to affect subsequent stimulation was also examined. Surprisingly, the response to PRL was inhibited by prior exposure for less than 3 h to either PRL or GH and disappeared with a longer pretreatment with either hormone. Similar to other tissues, the response to GH was inhibited by any length of prior exposure to GH. However, pretreatment with PRL had no effect. These experiments are the first demonstration of the transient desensitization of the PRL receptor by either PRL or GH pretreatment in any tissue and the desensitization of GH stimulation in islet cells. These observations provide insight into the mechanisms that regulate the desensitization of these receptors and, more importantly, allow the long-term activation of STAT5B by the PRL receptor. These results may apply to other members of the cytokine superfamily of receptors. We also demonstrate that the increase in islet cell proliferation required continuous stimulation with PRL, whereas the smaller effect with GH occurred with either continuous or pulsatile stimulation. In summary, this study demonstrates that islets are sensitive to the temporal pattern of stimulation by these hormones and provides a new basis for understanding their physiological roles in the regulation of islet function.
THE MOST POTENT growth factors that have been identified for pancreatic β-cells of the pancreas are prolactin (PRL) and GH secreted by the pituitary and the closely related placental lactogen (PL) secreted by the placenta during pregnancy (1–3). These hormones have been shown to stimulate insulin gene transcription, biosynthesis, and secretion, and increase β-cell proliferation. The physiological significance of these hormones is mainly indicated by the increased insulin secretion and islet mass during pregnancy in both rodents (4, 5) and humans (6). In rats, these changes correlate with the onset of secretion of PL (4). In contrast, a deficiency in PRL and GH in hypopituitary dwarf mice (5, 7) and hypophysectomized rats (8) is associated with impaired insulin secretion and reduced β-cell mass. The recent generation of PRL receptor-deficient mice demonstrated a similar impairment of insulin secretion and reduced islet mass (9). In contrast, increased insulin secretion and islet mass occurs in rats with a transplantable mammosomatotropic tumor that secretes both PRL and GH (10) and in transgenic mice with an overexpression of GH (5) or PL (11).
These effects are mediated through the PRL and GH receptors expressed in both insulinoma cell lines and primary β-cells (12–15). A distinct PL receptor has not been identified; rather, PL binds to the PRL receptor and, in primates, to the GH receptor in the presence of high Zn2+ (16). To determine whether these hormones have unique or shared effects on β-cells, we examined the effects of the homologous PRL, GH, and PL on islets of Langerhans isolated from several different species (17). This was done because their high degree of structural relatedness allows these hormones to exhibit activities in heterologous systems (i.e. hormones and islets from different species) that do not correspond to their effects in homologous systems (i.e. hormones and islets from the same species) (18). In each species, the homologous hormones that bind to the PRL receptor (i.e. PRL and PL) induced marked increases in insulin secretion and β-cell proliferation. In contrast, the homologous GHs had smaller or no effects. This observation was surprising considering the number of studies demonstrating large effects of GH on islet function (19, 20). However, many of these studies are misleading because rat islets were used with heterologous human GH that binds equally well to both the rat PRL and GH receptors.
This difference in effectiveness of the homologous hormones on islet function is surprising because the PRL and GH receptors are closely related members of the cytokine superfamily of receptors and have been shown to activate similar signaling cascades (21). Ligand binding induces receptor dimerization, the activation of the receptor-associated Janus kinase 2 (JAK2), and the tyrosine phosphorylation of signal transducers and activators of transcription (STAT) proteins. The activated STAT proteins dimerize and translocate to the nucleus where they bind specific DNA elements known as interferon-γ activated sequence sites and activate transcription. Although these receptors can activate multiple STAT proteins, the generation of STAT5A and STAT5B knockout mice showed that these STAT5 isoforms have essential roles in the biological actions of PRL and GH (22, 23). Moreover, the ability of human GH to stimulate the proliferation in the Ins-1 β-cell line is abolished by expression of a dominant negative STAT5 mutant (24). In pancreatic β-cells, STAT5 activation has been shown to stimulate the expression of the genes for the PRL receptor (14), glucokinase (25–27), cyclin D2 (24, 28), and insulin (29).
Because STAT5 activation has a central role in PRL- and GH-mediated signal transduction in β-cells, the relative potencies of PRL and GH on insulin secretion and β-cell proliferation may depend on their ability to maintain STAT5 activation. Recently, we have shown that continuous PRL stimulation induces a transient activation of STAT5A and a biphasic activation of STAT5B in β-cells (15). Before this study, the kinetics of STAT5A and STAT5B activation in β-cells by continuous GH stimulation were unknown. In addition, the secretion of GH by the pituitary in a pulsatile manner has been shown in all mammalian species examined so far (30). The activation of STAT5B in hepatocytes has been shown to be sensitive to this temporal pattern of GH stimulation (31–34). Although continuous stimulation with GH induces desensitization (33), pulsatile GH is more effective in maintaining the activation of STAT5B (35) and the expression of male-specific genes in the liver (36). If a similar sensitivity to the temporal pattern of stimulation by GH occurs in islets, it may be responsible for the difference in effectiveness of PRL and GH in the up-regulation of islet function.
The aim of the present study was to characterize the activation of the JAK2/STAT5 pathway by continuous and pulsatile stimulation with homologous rat PRL and GH in the insulin-producing Ins-1 β-cell line and intact pancreatic islets of Langerhans. Although most studies demonstrate the activation of STAT5 by examining the tyrosine phosphorylation, we have also used an immunohistochemical method to quantify the translocation of STAT5 from the cytoplasm to the nucleus after stimulation (15, 37). The major advantages of this approach are that it can be used to examine a mixed population of cells, can be used with tissue sections to examine the activation of STAT5 in vivo, and is very sensitive and thus able to detect changes in STAT5 translocation as small as 2% of the total cell content (15). These experiments suggest that the difference in response of islets to activation of these closely related receptors is due to the desensitization of islet cells to GH. This new information provides a basis for understanding the direct effects of these hormones on islet function and the difficulties of interpreting the observations from many previous studies.
Materials and Methods
Hormones
The rat PRL and GH used in this study were obtained from the National Hormone and Pituitary Program (Dr. A. F. Parlow, Harbor-University of California Los Angeles Medical Center, Torrance, CA) of the National Institute of Diabetes and Digestive and Kidney Diseases (Baltimore, MD). In most experiments, we used concentrations that have been shown to be slightly higher than those needed to induce their maximum effects on insulin secretion and cell proliferation for Ins-1 cells (200 ng/ml) (37) and isolated rat islets (500 ng/ml) (38).
Ins-1 cell culture
The rat insulinoma cell line Ins-1 (39) between passages 75 and 83 were cultured in complete medium consisting of RPMI 1640 medium with 10 mm glucose supplemented with 10 mm HEPES, 10% heat-inactivated fetal bovine serum, 1 mm pyruvate, 50 μm β-mercaptoethanol, 100 U/ml penicillin, 100 μg/ml streptomycin, and 250 ng/ml amphotericin B. Before an experiment, the monolayers were cultured for an additional 48 h in a defined serum-free medium (15, 37) consisting of RPMI 1640 medium with 10 mm glucose supplemented with 0.1% human serum albumin, 10 mm HEPES, 10 μg/ml human transferrin, 0.1 nm triiodothyronine, 50 μm ethanolamine, 50 μm phosphoethanolamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 250 ng/ml amphotericin B. All cells were cultured at 37 C in a humidified atmosphere of 5% CO2 in air.
Immunohistochemistry for STAT5 in Ins-1 cells
The Ins-1 cells (approximately 3 × 105) were cultured on 22-mm2 glass coverslips for 24–48 h in complete medium. Before the addition of hormones, the cells were cultured for an additional 48 h in the defined serum-free medium and were less than 25% confluence to facilitate the visualization of individual cells. The cells were briefly rinsed in PBS and fixed in 4% paraformaldehyde in 0.1 m phosphate buffer (pH 7.0) for 30 min at room temperature. After several brief rinses in PBS, the cells were permeabilized by incubation in Sorensen’s phosphate buffer containing 0.1% Triton X-100 for 20 min. The cells were then incubated overnight at 4 C with rabbit polyclonal antibodies against the carboxyl termini of mouse STAT5A (SC-1081) and STAT5B (SC-836) obtained from Santa Cruz Biotechnology (Santa Cruz, CA). The primary antibodies were diluted in PBS containing 0.3% Triton X-100, 1% normal donkey serum, and 1% BSA. The optimal dilution for each lot of these antibodies was determined by finding the dilution that showed the largest nuclear translocation of STAT5 without a substantial decrease in the overall intensity of staining (15). The optimal dilution for the individual lots of these antibodies varied from 1:400 (or 0.5 μg IgG/ml) to 1:1600 (or 0.125 μg IgG/ml). After several rinses in Sorensen’s phosphate buffer containing 0.1% Triton X-100 (4 × 30 min), the cells were then incubated for 4 h at 4 C with a 1:600 dilution of cyanine 3.18-conjugated donkey antirabbit IgG (Jackson Immunoresearch Laboratories, West Grove, PA) in PBS containing 0.3% Triton X-100, 1% normal donkey serum, and 1% BSA. After several rinses in Sorensen’s phosphate buffer containing 0.1% Triton X-100 (4 × 30 min), the monolayers were mounted in a 90% glycerol-10% PBS (pH 9.5) medium containing the antifade agent p-phenylenediamine.
Rat islet isolation and culture
Pancreatic islets were isolated from 3- to 5-d-old Sprague Dawley rats (Sasco, Omaha, NE) by a nonenzymatic culture method described previously (17). After this initial culture, groups of 30 islets were cultured free floating in 24-well plates (Costar, Cambridge, MA) in 2 ml of RPMI 1640 with 10 mm glucose supplemented with 10% horse serum, 25 mm HEPES, and 1% penicillin-streptomycin-fungizone antibiotic-antimycotic (Sigma, St. Louis, MO). Before the addition of hormones, the islets were then cultured for 48 h in a low-serum medium consisting of RPMI 1640 medium with 10 mm glucose supplemented with 1% horse serum, 25 mm HEPES, and 1% penicillin-streptomycin-fungizone antibiotic-antimycotic. The effect of PRL and GH on islet function was examined by culturing islets for 6 d with 0–1000 ng/ml PRL and measuring the release of insulin into the culture media and islet cell proliferation by adding 10 μm 5-bromo-2′-deoxyuridine during the final 24 h of culture (38). All islets were cultured at 37 C in a humidified atmosphere of 5% CO2 in air. Animal studies were conducted in accordance with the institutional laboratory animal care and use committee guidelines, and the protocols were approved by the institutional review board.
Immunohistochemistry for STAT5 in rat islets
The localization of STAT5A and STAT5B in isolated rat islets was done using the previously described procedure for Ins-1 cells with the following modifications. To identify the different islet cell types, islets were double- or triple-labeled by incubation with a 1:200 dilution of guinea pig antiinsulin antibody, a 1:1600 dilution of mouse monoclonal antiglucagon antibody (Sigma), or a 1:400 dilution of mouse monoclonal antisomatostatin (Sigma). The secondary antibodies also included a 1:600 dilution of cyanine 5.18-conjugated donkey antiguinea pig IgG (Jackson Immunoresearch Laboratories). Because of the large size of the intact islets, they were also incubated with the secondary antibodies overnight at 4 C. Glass beads of 50- to 100-μm maximum diameter were included in the mounting media to support the coverslips and prevent excessive deformation of the islets.
Confocal microscopy and image analysis
The immunostained specimens were examined with a Bio-Rad Lasersharp 1024 Confocal Imaging System (Bio-Rad Laboratories, Hercules, CA) (40) mounted on an Olympus AX70 microscope equipped for epifluorescence (Olympus, Lake Success, NY). The fluorescein isothiocyanate and cyanine 3.18 fluorophores were imaged using 488-nm excitation and a green bandpass emission filter (i.e. 505–540 nm) and 568-nm excitation and a red bandpass emission filter (i.e. 664–696 nm), respectively. The subcellular distribution of STAT5 was quantified using version 5.03 of the Confocal Assistant program written by T.C.B. (available from T. C. Brelje; University of Minnesota Medical School, Minneapolis, MN). This software was modified to allow the segmentation of individual cells and to store these results in data files. For each cell, the image passing through the maximum diameter of the nucleus was determined, and the borders of the cell and its nucleus were traced. The average fluorescence intensity of these regions was evaluated and used to calculate a nuclear to cytoplasmic intensity ratio for each cell. To aid the user, the outlines for each cell were drawn on this image and also on adjacent images in a different color to show which cells had already been analyzed.
For the Ins-1 cells, between 10 and 20 optical sections with a pixel size of 0.22 μm were acquired at 1.0-μm intervals along the z-axis through the monolayers using a ×60 lens (numerical aperture, 1.4). A total of at least 100 cells from three separate fields-of-view were analyzed for each experimental condition.
For the isolated rat islets, 11 optical sections were acquired at 2.0-μm intervals along the z-axis across the top of each islet using a ×40 lens (numerical aperture, 0.95). Five islets were analyzed for each experimental condition, and each series of images contained between 100 and 150 β-cells, 40 and 60 α-cells, and 10 and 20 δ-cells.
Western blot analysis
The Ins-1 cells (approximately 5 × 106) were cultured in 100-mm tissue culture dishes for 24–48 h in complete medium. Before an experiment, the cells were cultured for an additional 24 h in a defined serum-free medium and were at 50–60% confluence. The cells were then incubated in the presence or absence of 200 ng/ml PRL as indicated. The cells were washed twice with ice-cold PBS containing 1 mm sodium vanadate and lysed in 400 μl of lysis buffer containing 50 mm Tris (pH 8.0), 150 mm NaCl, 10 μg/ml BSA, 1 mm sodium vanadate, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% sodium dodecylsulfate, and 25 mm sodium fluoride supplemented with Complete Protease Inhibitor Cocktail (Roche Molecular Biochemicals, Indianapolis, IN) for 1 h on a rocking bench at 4 C. The lysates were then removed and cleared by centrifugation at 15,000 × g for 15 min. The loading volumes were determined by measuring the protein concentration of each lysate using 10-μl aliquots and the BCA assay (Pierce, Rockford, IL). A volume of lysate equivalent to approximately 5 × 106 cells was incubated with 5 μl of the immunoprecipitating antibody for 4 h at 4 C. This was followed by the addition of 25 μl of protein A-Sepharose beads (CL-4B; Pharmacia Biotech, Alameda, CA) and further incubation for 4–18 h at 4 C. The pellets of Sepharose beads were washed twice with lysis buffer, twice with 10 mm Tris buffer (pH 8.0) containing 150 mm NaCl, 0.1% BSA, and 0.1% Triton X-100, once with 10 mm Tris buffer (pH 8.0) containing 150 mm NaCl, and once with 50 mm Tris buffer (pH 6.8). The pellets were then incubated in electrophoresis treatment buffer for 5 min at 90 C. These samples were electrophoresed in 8% acrylamide gels and electroblotted onto Immobilon-P membranes (Millipore, Bedford, MA). The blots were temporarily stained with Ponceau red to assure protein transfer and to determine the location of molecular weight standards. The blots were incubated with 1% BSA for 4–18 h at 4 C to block nonspecific staining, then they were incubated with a 1:2000 dilution of the primary (i.e. probing) antibody overnight at 4 C, and finally, they were incubated with a 1:30,000 dilution of either alkaline phosphatase-conjugated donkey antirabbit or antimouse IgGs (Jackson ImmunoResearch Laboratories) for an additional 45 min at room temperature. A mouse monoclonal antibody against phosphotyrosine (4G10; Upstate Biotechnology, Lake Placid, NY) was used to determine the extent of tyrosine phosphorylation of the immunoprecipitated protein. The blots were then processed for the probing antibodies using the chemiluminescent substrate CSPD (Tropix, Bedford, MA) and BioMax-L film (Eastman Kodak, Rochester, NY). Quantitative densitometry was done with a Bio-Rad GS-700 imaging densitometer and the Molecular Analyst software (Bio-Rad). The volume density of the chemiluminescent bands was calculated as OD × mm2 after background correction. To confirm equivalent loading of each lane, the blots were stripped of the bound antibodies and reprobed using the immunoprecipitating antibody.
Data analysis and presentation of results
All results are expressed as the mean ± sd of the number of observations. Statistical differences between the means were assessed by either the Student’s t test or Student Neuman-Keuls one-way ANOVA for multiple comparisons.
Results
Kinetics of STAT5 translocation in Ins-1 cells
The effect of these hormones on the subcellular localization of STAT5 was examined in Ins-1 cells exposed to 200 ng/ml PRL or GH for 0–6 h. STAT5A and STAT5B localization was then determined by indirect immunohistochemistry (Fig. 1). In unstimulated cells, both were detected throughout the cell as a fine punctuate pattern with slightly higher intensity in the nucleus compared with the cytoplasm. After PRL or GH stimulation, there was a marked increase in STAT5 immunoreactivity within the nucleus. As previously described (15), the activation of STAT5 was monitored by using the average fluorescence intensity of the nucleus and cytoplasm to calculate a nuclear to cytoplasmic intensity ratio for individual cells (Fig. 2). The nuclear translocation of both STAT5A and STAT5B peaked within 30 min of exposure to either hormone, with the relative increase for STAT5B being noticeably greater than for STAT5A. After this initial peak, their responses subsequently diverged. For PRL stimulation, the translocation of STAT5A decreased to a level slightly above that observed before stimulation, whereas the translocation of STAT5B was maintained at the same level as the initial peak. In contrast, the translocation of both STAT5A and STAT5B decreased to control levels with continued GH stimulation. These differences suggest that activation of the PRL receptor, but not the GH receptor, can maintain the nuclear translocation of STAT5 in β-cells.
Kinetics of tyrosine phosphorylation of JAK2 and STAT5 in Ins-1 cells
The time course of the PRL- or GH-induced tyrosine phosphorylation of JAK2, STAT5A, and STAT5B in Ins-1 cells was examined (Fig. 3, A–C). In unstimulated cells, it was not possible to detect the tyrosine phosphorylation of these proteins. After PRL stimulation, the tyrosine phosphorylation of each protein was biphasic, with an initial peak within 30 min, a nadir between 1 and 3 h, and another increase after 4 h. The relatively similar profiles for STAT5A and STAT5B were surprising considering the previously observed differences in their nuclear translocation (Figs. 1 and 2). After GH stimulation, the tyrosine phosphorylation of each protein peaked within 1 h at higher levels than with PRL and then decreased to much lower levels. The profiles for STAT5A and STAT5B were similar to their previously observed changes in STAT5 translocation (Fig. 2). Although the total protein levels for STAT5A and STAT5B were unchanged, there was a marked decline in the protein levels of JAK2 after stimulation for 1 h with either hormone (Fig. 3D). In experiments of up to 48 h, this decline continued to a level around 40% of that observed at the onset of the experiment (data not shown). These observations demonstrate that PRL can maintain the tyrosine phosphorylation of JAK2, STAT5A, and STAT5B in β-cells, whereas the response to GH is transient. This difference is most likely responsible for the observed differences in STAT5 translocation.
STAT5B translocation in β-, α-, and δ-cells of rat islets
Because PRL stimulates the nuclear translocation of STAT5 only in the β-cells of rat islets (15), we examined whether GH had a similar cellular specificity. Rat islets were exposed to 500 ng/ml PRL, 500 ng/ml GH, or 10% fetal bovine serum for 30 min and then double-labeled with antibodies to STAT5B and either insulin, glucagon, or somatostatin to identify different islet cell types. PRL stimulated the nuclear translocation of STAT5B only in the insulin-containing β-cells and not in the glucagon-containing α-cells or the somatostatin-containing δ-cells (Fig. 4). In contrast, GH and fetal bovine serum stimulated the nuclear translocation of STAT5B in all islet cells (Fig. 4). These observations demonstrate that all islet cells have a functional STAT5 signaling pathway but that only β-cells are PRL responsive.
Kinetics of STAT5B translocation in rat islets
The time course for the nuclear translocation of STAT5B in rat islets was also examined. The nuclear to cytoplasmic intensity ratio was determined for both β- and α-cells in rat islets after incubation with 500 ng/ml PRL or GH for 0–48 h. PRL induced a biphasic translocation of STAT5B in β-cells with an initial peak within the first hour, followed by a nadir between 1 and 3 h, and another peak after 4 h of stimulation (Fig. 5, top left). This biphasic pattern was more pronounced than that observed in the Ins-1 cells (Fig. 2). In contrast, GH induced a monophasic translocation of STAT5B in β-cells with a larger initial peak within the first hour followed by a decrease to a level slightly above the controls (Fig. 5, top left). Consequently, PRL induced a larger increase in STAT5B translocation in β-cells than GH after 3 h of stimulation. In α-cells, GH stimulated a transient translocation of STAT5B similar in profile to that observed in β-cells (Fig. 5, top right). The combination of both hormones resulted in a biphasic translocation of STAT5B similar to the sum of the individual responses (Fig. 5, bottom left). These observations suggest there are substantial differences in the activation of STAT5B by the closely related PRL and GH receptors in normal β-cells.
Dose response for STAT5B translocation in rat islets
Because the maximum extent of translocation in β-cells was higher with GH compared with PRL (Fig. 5), the effectiveness of different concentrations of these hormones to induce the nuclear translocation of STAT5B was also examined. The nuclear to cytoplasmic intensity ratio was determined for β- and α-cells in rat islets after incubation with 0–5000 ng/ml PRL or GH for 30 min (Fig. 6, left). PRL induced the nuclear translocation of STAT5B in β-cells at concentrations as low as 50 ng/ml, with a maximum effect observed above 500 ng/ml. GH induced the nuclear translocation of STAT5B in both β- and α-cells at concentrations as low as 10 ng/ml, with a maximum effect observed above 100 ng/ml GH. Although similar responses were observed in both β- and α-cells at the lower GH concentrations, the maximum extent of translocation was higher in β-cells compared with α-cells. Similarly, the maximum extent of translocation in β-cells was higher with GH compared with PRL. Therefore, 500 ng/ml of either hormone was sufficient to induce their maximum effects on STAT5B translocation.
Because the effects of PRL in Ins-1 cells are abolished by the expression of a dominant negative STAT5 mutant (24), it was interesting to compare these dose response curves to those for their effects on islet function. Rat islets were cultured with 0–1000 ng/ml PRL or GH for 6 d, and insulin secretion (Fig. 6, middle) and β-cell proliferation (Fig. 6, right) were examined. Although both dose-response curves for PRL are remarkably similar to those for STAT5B translocation, this is not the case for GH. This suggests that the larger but transient activation of STAT5B with GH does not compensate for its inability to maintain the activation of STAT5B with prolonged stimulation. Therefore, these results are consistent with the observation that long-term activation of STAT5B is a key event in the up-regulation of islet function.
Homologous desensitization of STAT5B translocation
Several studies have demonstrated that the activation of STAT5B in hepatocytes is sensitive to the temporal pattern of GH stimulation (31, 34, 35). Desensitization to GH occurs within the first 2 h, and full responsiveness to succeeding GH stimulation requires a minimum of 2.5 h without GH (31). Because this desensitization and resensitization are critical processes regulating GH actions on target tissues, we investigated whether a similar situation exists with PRL and/or GH in islets.
First, we examined whether pretreatment with the homologous hormone can induce desensitization. Islets were pretreated with 500 ng/ml of PRL or GH for 0–6 h, washed in hormone-free media for 2 h, and then restimulated with 500 ng/ml of the same hormone for 30 min. In islets stimulated with PRL, STAT5B translocation in β-cells was decreased with a pretreatment of less than 3 h (Fig. 7). This inhibition disappeared with a longer pretreatment. In islets stimulated with GH, STAT5B translocation in both β- and α-cells was decreased 40–60% with any length of pretreatment (Fig. 7). Thus, pretreatment with either hormone can desensitize islets to subsequent stimulation. However, this inhibition is transient with PRL, whereas any length of prior exposure to GH is sufficient to induce desensitization.
Second, the time-dependent recovery from this desensitization was examined. Islets were pretreated with 500 ng/ml of PRL or GH for 1 h, washed in hormone-free media for 1–6 h, and then restimulated with 500 ng/ml of the same hormone for 30 min. In islets stimulated with PRL, STAT5B translocation in β-cells was decreased with a wash interval of 1 h (Fig. 8). This inhibition was fully recovered after 3 h in hormone-free media. In islets stimulated with GH, STAT5B translocation in both β- and α-cells was inhibited with a wash interval of 1 h (Fig. 8). Although this inhibition was also reduced during the first 3 h in hormone-free media, full recovery was never achieved. Thus, the time required for recovery from the largest component of this desensitization is similar for both hormones. It is unclear whether the failure to fully recover responsiveness to GH is due to an exaggerated initial response after prolonged culture in hormone-free media or the presence of additional inhibitory mechanisms that affect GH alone.
Heterologous desensitization of STAT5B translocation
Because the PRL and GH receptors are structurally related and appear to induce similar signaling pathways, we examined whether pretreatment with PRL and/or GH can similarly induce desensitization to the other hormone in islets.
First, we examined whether pretreatment with GH can induce the heterologous desensitization of PRL-induced STAT5B activation. Islets were pretreated with 500 ng/ml PRL, GH, or the combination of both PRL and GH for either 1 or 6 h, washed for 1 h in hormone-free media, and then restimulated with 500 ng/ml PRL for 30 min. As previously shown, STAT5B translocation in β-cells was decreased by pretreatment with PRL for 1 h (Fig. 9). A similar decrease was observed after pretreatment with GH for 1 h. The combination of PRL and GH was no more effective than GH alone. This inhibition was absent for both hormones with a longer pretreatment of 6 h (Fig. 9). Thus, activation of the PRL receptor in β-cells can be transiently desensitized by either homologous stimulation by PRL itself or heterologous stimulation by GH.
Second, we examined whether pretreatment with PRL can induce the heterologous desensitization of GH-induced STAT5B activation. Islets were pretreated with 500 ng/ml PRL, GH, or the combination of both PRL and GH for either 1 or 6 h, washed for 1 h in hormone-free media, and then restimulated with 500 ng/ml GH for 30 min. As previously shown, STAT5B translocation in both β- and α-cells was decreased by pretreatment with GH for 1 h (Fig. 10). However, pretreatment with PRL for 1 h had no effect on the response in β-cells. Surprisingly, pretreatment with PRL led to an increase in STAT5B translocation in α-cells even though they were unresponsive to PRL. Similar results were observed with a longer pretreatment of 6 h (Fig. 10). Thus, activation of the GH receptor in β- and α-cells can be desensitized by homologous stimulation by GH but not by heterologous stimulation by PRL.
The absence of the heterologous desensitization of GH stimulation by pretreatment with PRL was unexpected. Because the extent of STAT5B translocation with 500 ng/ml GH is larger than with 500 ng/ml PRL, one possibility is that stimulation with GH is able to overcome any inhibition induced by pretreatment with PRL. For this reason, we examined whether desensitization could occur with concentrations of PRL and GH that induce similar levels of STAT5B activation in β-cells. Islets were pretreated with 500 ng/ml PRL or 70 ng/ml GH for 1 h, washed for 1 h in hormone-free media, and then restimulated with 500 ng/ml PRL or 70 ng/ml GH for 30 min. In the absence of pretreatments, these concentrations of hormones induced similar levels of STAT5B translocation in β-cells (Fig. 11). The homologous desensitization of GH with this lower concentration was approximately half of that observed with 500 ng/ml GH in both β- and α-cells. However, pretreatment with PRL was still unable to induce the heterologous desensitization of GH in β-cells.
β-Cell proliferation induced by continuous vs. pulse stimulation
Because the biological actions of GH in other tissues have been shown to be dependent on whether GH is secreted in a pulsatile or continuous fashion (30), it was of interest to determine whether a similar situation exists with islets. Islets were exposed to 500 ng/ml PRL, GH, or the combination of PRL and GH for 48 h either continuously or pulsed for 1 h with hormone(s) every 4 h (Fig. 12). Continuous stimulation with PRL led to a 2-fold increase in islet cell proliferation, whereas this effect was lost with pulsatile stimulation. In contrast, both continuous and pulsatile stimulation with GH led to much smaller increase (30%) in islet cell proliferation. Unlike in the liver, pulsatile stimulation with GH was not more effective than continuous stimulation. As expected from the results with the individual hormones, the combination of PRL and GH was similar to PRL alone with continuous stimulation and to GH alone with pulsatile stimulation.
Discussion
This study demonstrates that there are marked differences in the activation of STAT5 by PRL and GH in the Ins-1 β-cell line and normal rat islets of Langerhans. We monitored the activation of STAT5 using an immunohistochemical approach that examines its subcellular distribution within individual cells (15, 37). The reciprocal changes in STAT5 immunoreactivity detected within the nucleus and the cytoplasm after stimulation are consistent with the nuclear translocation of STAT5. Unlike Western blots prepared from tissue extracts, this method can be used to investigate the activation of STAT5 in a heterologous population of cells. We were able to demonstrate that only normal β-cells, and not α- or δ-cells, are responsive to PRL. In contrast, GH induced the nuclear translocation of STAT5B in all islet cells examined (i.e. α-, β-, and δ-cells). Furthermore, GH only transiently activates STAT5B and is subsequently desensitized to further stimulation. This desensitization to GH is similar to that previously described for the activation of STAT5B in the liver (31–33). Surprisingly, the response to PRL was inhibited by prior exposure for less than 3 h to either PRL or GH and disappeared with a longer pretreatment with either hormone. Similar to other tissues, the response to GH was inhibited by any length of prior exposure to GH. However, pretreatment with PRL had no effect. These experiments are the first demonstration of the transient desensitization of the PRL receptor by either PRL or GH pretreatment in any tissue and the desensitization of GH stimulation in islet cells. These observations suggest that continuous stimulation with PRL is considerably more effective than GH in up-regulating islet function because of the ability of PRL to maintain long-term STAT5B activation in β-cells.
Overall, the characteristics of STAT5 activation were similar in the transformed Ins-1 cells and the β-cells of isolated islets in response to PRL or GH stimulation. This result was surprising because both hormones are equally effective in stimulating the growth of Ins-1 cells (37). However, the extreme sensitivity of the Ins-1 cells to PRL (ED50 of 5–10 ng/ml) (37) compared with normal islets (ED50 of 200 ng/ml; Fig. 6) suggests that the Ins-1 cells are unusually sensitive to the low levels of STAT5 activation. Unlike primary β-cells, the low level of STAT5B activation detected after the initial peak of GH stimulation appears to be sufficient to increase cell division in the more rapidly growing Ins-1 cells. Although the kinetics for the biphasic activation of STAT5B by PRL are similar in Ins-1 cells (Fig. 3) and the β-cells of isolated islets (Fig. 5), the transient activation of STAT5B by GH lasted for 6 h in Ins-1 cells compared with less than 2 h in normal islets. This longer activation may also be related to the less effective desensitization of GH stimulation by pretreatment with GH in Ins-1 cells (our unpublished observations). Because of these differences, most of our experiments were done with isolated islets.
Our experiments demonstrate that the kinetics of STAT5 activation in β-cells by continuous stimulation with PRL or GH is different. This desensitization to GH is similar to that previously reported in hepatocytes (31, 33, 34). It is still controversial whether this GH-induced desensitization of GH signaling is the result of postreceptor signaling pathways (32, 33, 41, 42) or the down-regulation of GH receptor levels (34, 43). In the present study, the presence of parallel changes in the tyrosine phosphorylation of JAK2 and STAT5 suggests that the extent of JAK2 activation is responsible for this difference between PRL and GH. It has been proposed that a GH-inducible, labile inhibitor and/or phosphatase may regulate this process of GH desensitization in hepatocytes (32, 33, 44). However, the additive effects of both hormones on STAT5B activation in β-cells suggest that this process must be specific for the deactivation of JAK2 by the GH receptor but not the PRL receptor. A similar situation has been reported in IM-9 lymphocytes in which the activation of JAK2 by interferon-γ (IFN-γ) is unaffected by desensitization to GH (45).
Alternatively, if the GH-induced desensitization of its own pathway is achieved by down-regulating the GH receptor, it is simple to envision how the response to PRL is preserved. Both receptors have short half-lives on the surface of cells (< 1 h) and require ongoing protein synthesis to maintain the population of receptors on the cell surface (46, 47). Because greater than 75% of the PRL (48) or GH (49) receptors on the cell surface are internalized within 30 min of stimulation, the long-term rate of formation of activated receptor/JAK2 complexes may be limited by the replenishment of the cell surface receptors. Recently, evidence for the down-regulation of GH receptors being involved in the GH-induced homologous desensitization in rat hepatoma cells has been reported (34). This study detected a 70% reduction in the total cellular levels of the GH receptor after GH stimulation. Moreover, the recovery of GH responsiveness was also correlated with the time-dependent recovery of GH receptor levels. Although similar studies examining the total cellular levels of the GH receptor in islets have not been reported, both PRL and GH are able to decrease the expression of GH receptor mRNA in Ins-1 cells after 4 h of stimulation (13). The inability of PRL pretreatment to induce the heterologous desensitization of GH in our experiments suggests that this decrease in mRNA may not be sufficient to explain the desensitization to GH. However, the decrease in GH receptors may be much lower with GH stimulation because it also promotes the degradation of both ligand-bound internalized receptors and receptors in intracellular compartments (50). In contrast, both PRL and GH increase the expression of PRL receptor mRNA in Ins-1 cells and islets (12, 13). It has been shown that this transcriptional activation occurs through a STAT5 binding site in the promoter of the PRL receptor gene (14). The larger increase in PRL receptor mRNA in rat islets after 24 h with PRL compared with GH most likely reflects its ability to maintain a higher level of STAT5 activation. These observations suggest that the STAT5B-induced increase in the expression of the PRL receptor after continuous PRL stimulation is an essential component in its ability to maintain the long-term activation of STAT5B.
The experiments examining the homologous and heterologous desensitization of STAT5B activation by PRL and GH in β-cells also imply that postreceptor inhibitory signaling pathways may be involved in the desensitization of these receptors. The transient desensitization of STAT5B activation by PRL with pretreatment with either PRL itself or GH suggests that both receptors induce common inhibitory mechanisms in the first 3 h of stimulation. The suppressor of cytokine signaling (SOCS) and cytokine-inducible SH2-domain (CIS) proteins have been identified as feedback inhibitors of cytokine receptor signaling through a STAT-dependent transcriptional mechanism (51, 52). Although examined in different cell types, PRL and GH induce similar expression profiles of these proteins, with a transient increase in SOCS1 and SOCS3 peaking in the first hour of stimulation followed by a prolonged increase in SOCS2 and CIS. These proteins appear to have similar effects on the activation of STAT5 by both the PRL and GH receptors in cotransfection experiments (51, 52). They inhibit cytokine signaling by multiple mechanisms involving the binding to phosphotyrosine residues of JAK2 (SOCS1) or the cytoplasmic tail of these receptors (SOCS3 and CIS). Moreover, STAT5 activation in Ins-1 cells by human GH (which binds to both rat PRL and GH receptors) has been shown to be inhibited by the expression of SOCS3 under the control of an inducible promoter (53). SOCS1 and SOCS3 have also been proposed to mediate the heterologous inhibition of STAT1 activation by IFN-γ in Ins-1 cells by pretreatment with GH (54). Unfortunately, it was not determined whether the activation of STAT5 by GH could be similarly inhibited by pretreatment with IFN-γ. These observations suggest that the transient expression of SOCS1 and/or SOCS3 accounts for the biphasic activation of STAT5B with continuous PRL stimulation and the transient desensitization of PRL after pretreatment with either PRL or GH in β-cells. The inability of pretreatment with PRL to inhibit STAT5B activation by GH may be due to differences in their protein levels induced by these receptors, the sensitivity of these receptors to their inhibitory effects, or the requirement for additional factors only induced by GH. Interestingly, a similar situation has been reported in the pro-B lymphocyte cell line Ba/F3 in which both IL-3 and IL-11 induce similar amounts of SOCS-3, but only IL-3 is able to block the activation of STAT3 by IL-11 (55).
Our failure to observe an inhibition of STAT5B activation in β-cells by PRL after pretreatment with either PRL or GH for more than 3 h also suggests that the down-regulation of the GH receptor is responsible for the long-term desensitization of GH signaling. However, the increase in CIS expression after the first hour of stimulation has been proposed as a key mediator of the desensitization of GH signaling with continuous stimulation (42). Although initial studies failed to detect an inhibition of GH signaling by elevated CIS expression (52), it was subsequently shown to be a potent inhibitor when expressed at higher levels than STAT5B (42). Because PRL and GH induce similar expression profiles of CIS (51, 52), PRL signaling would have to be less sensitive to the inhibitory effects of CIS to account for our failure to detect the inhibition of STAT5B activation in β-cells by PRL after pretreatment with either hormone for more than 3 h. However, overexpression of CIS has also been shown to inhibit PRL signaling (56). The differences in long-term STAT5B activation by the PRL and GH receptors presumably in the presence of CIS suggest that CIS has a different role. CIS has been shown to inhibit GH signaling by the following two distinct mechanisms: a partial inhibition that is decreased at elevated STAT5B levels and may involve competition between CIS and STAT5B for common phosphotyrosine-binding sites in the cytoplasmic tail of the GH receptor, and a time-dependent inhibition (not seen with SOCS1 or SOCS3) that involves proteasome action (42). In addition, the down-regulation of GH signaling with continuous stimulation could be prevented by the proteasome inhibitor MG132 or expression of a dominant-negative inhibitor of CIS activity (42). This observation suggests that CIS may be responsible for promoting the inactivation of the activated receptor/JAK2 complexes by both PRL and GH at normal levels of expression. Only at higher expression levels when it can compete with STAT5B binding to the phosphotyrosine residues of the receptors is CIS an effective inhibitor. The increased expression of CIS with prolonged PRL stimulation may also explain why the extent of STAT5B activation after 4 h of continuous PRL stimulation is not noticeably higher than the initial peak after 30 min even though the receptor levels should be higher.
These results suggest that the capabilities of the cytokine receptor superfamily of receptors may be more diverse than originally appreciated. Although the transient activation of STATs is widely acknowledged (21), our experiments demonstrate that receptors that induce a prolonged or transient activation of a STAT5B can coexist in the same cell. Evidence for a similar situation can be found in a study examining the expression of the IL-2 receptor in T-cell lymphoma P60 cells (57). In these cells, IL-2 induces the long-term STAT5 activation and increased cell surface expression of the IL-2 receptor. In contrast, erythropoietin only induces transient STAT5 activation with no change in IL-2 receptor levels. Similar to the PRL-induced increase in the expression of the PRL receptor, the increased expression of the IL-2 receptor occurs through a STAT5 binding site in the promoter of the IL-2 receptor gene (58). These observations suggest that transient vs. long-term activation of STATs may be a general property of the cytokine receptor superfamily and may be related to the ability of individual hormones to regulate the expression of their own receptor. This difference allows multiple receptors that use the same signaling pathway to coexist in the same cell and be sensitive to different temporal patterns of stimulation. Because the β-cell expresses other members of the cytokine receptor superfamily, it provides an excellent opportunity to investigate the interactions between these different receptors.
Besides providing information about the STAT5 activation in β-cells by PRL and GH receptors, these results have important implications for our views about the physiological roles of these hormones in the regulation of pancreatic islets. A tonic stimulation of islets by either PRL or GH is required for the normal responsiveness of islets to glucose stimulation. This effect is most likely related to the ability of STAT5 to stimulate the expression of the glucokinase (25–27) and insulin (29) in β-cells. The rapid desensitization of the GH response explains why its effects on islet function are so small compared with the effects with a continuous stimulation with PRL or PL. In all mammalian species examined so far, spontaneous episodes of GH secretion occur several times over a 24-h period (30). In the rat, the pituitary secretes GH in a pulsatile manner in males and in a nearly continuous fashion in females (30). The male-specific pattern of pulsatile GH secretion is more effective than continuous GH in promoting weight gain associated with long bone growth and the transcriptional regulation of sexually dimorphic genes in the liver (30, 59). The mechanism underlying this phenomenon is based on the desensitization of the response to GH with prolonged exposure. This desensitization to continuous GH results in levels of tyrosine phosphorylation of STAT5B that are approximately 10% of the maximal levels observed during the first 1 h of GH stimulation. The importance of the pulsatile activation of STAT5B in the liver is demonstrated in male mice with a knockout of the STAT5B gene in which their body size and sex-specific gene expression in the liver is comparable to females (59). However, we were unable to demonstrate that pulsatile GH was more effective than continuous GH in the stimulation of β-cell proliferation in rat islets. This difference between the liver and islets requires further study and may be related to the age of the animals. However, the absence of marked differences in insulin secretion between males and females in rats suggests that the low level of STAT5B activation with continuous GH (or from the normal circulating levels of PRL) is sufficient to maintain the normal glucose sensitivity of islets. This suggests that the effects of GH on islet function may be limited to those observed with continuous GH stimulation.
Although our experiments were done in rats, there is evidence to suggest that a similar difference between the effects of PRL and GH on pancreatic islets also occurs in humans. A chronic elevation of GH in humans is associated with an impairment of glucose tolerance and an increased incidence of diabetes (60, 61). This is most likely due to GH counteracting the insulin stimulation of glucose uptake in peripheral tissues. However, our experiments suggest that the desensitization of β-cells to the chronically elevated GH may also contribute to this situation. In contrast, a chronic elevation of PRL in humans is associated with enhanced insulin secretion and a slight decrease in fasting serum glucose concentrations (62, 63). These changes are consistent with the ability of PRL to maintain the long-term activation of STAT5B in β-cells and up-regulate islet function. Furthermore, the inability of long-term exposure to GH to inhibit PRL in our experiments may also be reflected in humans because an elevation in PRL also occurs in 25–50% of the patients with acromegaly (64). This may explain why an abnormal glucose tolerance is observed in 68% of human patients with acromegaly, but less than 15% of the patients are diabetic (65). It would be interesting to know whether there was an inverse relationship between an elevation of PRL and the occurrence of abnormal glucose tolerance in humans with acromegaly.
The most compelling evidence for lactogen activation of islets occurs during pregnancy. Pregnancy is characterized as a condition of elevated serum insulin levels, slightly lower blood glucose levels, and increased peripheral insulin resistance (66). To accommodate this increased demand for insulin and maintain normal glycemia, the islets must undergo long-term up-regulatory changes in function. Failure of this long-term adaptive process can lead to the development of gestational diabetes. The principal changes that occur in islets during pregnancy are an increase in glucose-stimulated insulin secretion and islet mass (4, 5). This lowering of the threshold for glucose-stimulated insulin secretion dramatically increases the amount of insulin secreted at normal serum glucose concentrations. The stimulation of β-cell proliferation may be important in maintaining a sufficient islet mass for this increased rate of insulin secretion. Although an elevation in serum glucose concentrations by itself could stimulate islet function, the resulting hyperglycemia would be detrimental to both the mother and the developing fetus. The physiological relevance of PL secretion has been questioned because placentas of many species do not produce a distinct PL (for example, rabbits, pigs, dogs, and cats) (67). However, an increase in serum levels of PRL may compensate for the absence of a distinct PL during pregnancy. An increase in serum levels of PRL during pregnancy has been shown in dogs (68) and humans (69, 70). The prolonged elevation of lactogens in the form of PL and/or PRL during gestation should support the long-term activation of STAT5b in β-cells. In contrast, the pulsatile secretion of PRL during lactation would be insufficient to maintain the up-regulation of islet function and may be part of the explanation of why they are in a lowered functional state during this period.
Acknowledgments
We thank Dr. A. F. Parlow from the National Hormone and Pituitary Program of the National Institute of Diabetes and Digestive and Kidney Diseases for providing the rPRL and rGH used in this study.
This work was supported by National Institutes of Health Grant DK33655.
Abbreviations:
- CIS,
Cytokine-inducible SH2-domain;
- IFN-γ,
interferon-γ;
- JAK2,
Janus kinase 2;
- PL,
placental lactogen;
- PRL,
prolactin;
- SOCS,
suppressor of cytokine signaling;
- STAT,
signal transducers and activators of transcription.
References