Abstract
This study examined whether an immunohistochemical method examining the subcellular localization of STAT5 could be used to characterize the activation of the JAK2/STAT5 pathway by prolactin (PRL) in intact cells or tissues. In the Ins-1 β-cell line, STAT5A and STAT5B were distributed almost equally in the cytoplasm and the nucleus in unstimulated cells. STAT5A was also detected along the border of cells and in the perinuclear region. After exposure to PRL, the redistribution from the cytoplasm to the nucleus was much higher for STAT5B compared to STAT5A. This translocation represented 12% of the STAT5A and 22% of the STAT5B originally located in the cytoplasm before stimulation. In isolated rat islets of Langerhans, PRL stimulated the nuclear translocation of both STAT5A and STAT5B only in β-cells. The expression of the PRL receptor only by β-cells was confirmed with a rabbit polyclonal antiserum raised against the rat PRL receptor. It was estimated that 4% of STAT5A and 9% of STAT5B originally located in the cytoplasm was translocated to the nucleus after stimulation. The presence of a functional JAK2/STAT5 signaling pathway in all islet cells was demonstrated by the nuclear translocation of STAT5B in all islet cells (i.e., α-, β-, and δ-cells) after stimulation with fetal calf serum. The nuclear translocation and tyrosine phosphorylation of STAT5B was biphasic, with an initial peak within 30 min, a nadir between 1 and 3 hr, and prolonged activation after 4 hr. In contrast, the tyrosine phosphorylation of STAT5A was also biphasic but its nuclear translocation peaked within 30 min and was then reduced to a level slightly above that observed before PRL stimulation. This method is able to detect changes in STAT5 activation as small as 2% of the total cell content. These observations demonstrate the utility of this approach for studying the activation of STAT5 in a mixed population of cells within tissues or organs. In addition, the dose response for the nuclear translocation of STAT5B in normal β-cells was similar to those for changes in proliferation and insulin secretion in isolated rat islets. Therefore, the subcellular localization can be used to monitor the activation of STAT5 and it may be a key event in the upregulation of the pancreatic islets of Langerhans during pregnancy.
The Prolactin (PRL) RECEPTOR belongs to the superfamily of cytokine receptors characterized by their ability to activate Janus kinases (JAKs) and the STAT (signal transducers and activators of transcription) transcription factors (Ihle 1996; Darnell 1997). In the unstimulated state, STATs are localized throughout the cell. On stimulation they are phosphorylated, which induces their dimerization and translocation into the cell nucleus, where they bind to specific DNA elements known as γ-activated sequence-like elements (GAS) and modulate the expression of target genes. The PRL receptor activates a subset of STAT proteins including STAT1, STAT3, STAT5A, and STAT5B (Bole-Feysot et al. 1998). Although the physiological significance of STAT1 and STAT3 activation by PRL is unclear, the generation of STAT5A and STAT5B knockout mice showed that these STAT5 isoforms have essential roles in the biological actions of PRL (Liu et al. 1997; Teglund et al. 1998).
Although most studies demonstrate the activation of STAT5 by examining their tyrosine phosphorylation, we examined whether an immunohistochemical (IHC) method could be used to quantify the redistribution of STAT5 from the cytoplasm to the nucleus. However, it is unknown whether the available antibodies are able to recognize their respective antigens with equal efficiency before and after activation in intact cells. If this is an unusual occurrence, an IHC approach would have several advantages: first, whether heterogeneity occurs in the response of similar cells could be examined; second, the response of individual cell types in a mixed population of cells could be investigated; and third, it could be used with tissue sections to examine the activation of STAT5 in vivo. Finally, this approach would allow the simultaneous measuring of cytoplasm to nuclear translocations of transcription factors within individual cells.
The aim of the present study was to characterize the PRL-induced activation of the JAK2/STAT5 pathway in the insulin-producing Ins-1 β-cell line and intact pancreatic islets of Langerhans. We have shown that prolonged culture of islets with hormones that bind to the PRL receptor result in increased insulin secretion and cell division (Parsons et al. 1992,1995; Brelje et al. 1993a,b,1994; Weinhaus et al. 1996). This response is particularly important during pregnancy, when placental lactogen and/or PRL, acting through the PRL receptor, are responsible for the up-regulation of islet function (Brelje and Sorenson 1997; Sorenson and Brelje 1997,2001). The activation of STAT5 mediates the expression of a number of tissue-specific genes, including those for insulin (Galsgaard et al. 1996), the PRL receptor (Galsgaard et al. 1999), glucokinase (Weinhaus et al. 2001), and cyclin D2 (Friedrichsen et al. 2001) in β-cells. However, it remains to be determined whether the characteristics of STAT5 activation in islets are consistent with the known changes in islet function.
Materials and Methods
Hormones
The rat PRL (NIDDK-rPRL-B-8-SIAFP) used in this study was obtained from the National Hormone and Pituitary Program (Dr. A. F. Parlow, Harbor-UCLA Medical Center, Torrance, CA) of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK; Baltimore, MD).
Ins-1 Cell Culture
The rat insulinoma cell line Ins-1 (Asfari et al. 1992,1995) between passages 75 and 83 was 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 hrs in a defined serum-free medium (Clark and Chick 1990) 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 37C 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 hr in complete medium. Before addition of hormones, the cells were cultured for an additional 48 hr in the defined serum-free medium and were at less than 25% confluence to facilitate 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 (RT). 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 4C 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% bovine serum albumin (PBS/T). The optimal dilution for each lot of these antibodies was determined by finding the dilution that showed the highest PRL-induced nuclear translocation without a substantial decrease in the overall intensity of staining. 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 (four times for 30 min), the cells were incubated for 4 hr at 4C with a 1:600 dilution of cyanine 3.18 (CY3)-conjugated donkey anti-rabbit IgG (Jackson Immunoresearch Laboratories; West Grove, PA) in PBS/T. After several rinses in Sorensen's phosphate buffer containing 0.1% Triton X-100 (four times for 30 min), the monolayers were mounted in a 90% glycerol/10% PBS (pH 9.5) medium containing the antifade agent p-phenylenediamine.
Double labeling for STAT5A localization utilized the above procedure except that the primary antibody also included a 1:800 dilution of mouse anti-nuclear transport factor p97 antibody (Affinity BioReagents; Golden, CO), a 1:400 dilution of mouse anti-transferrin receptor antibody (Pharmingen; San Diego, CA), a 1:400 dilution of mouse anti-β-COP antibody (Sigma; St Louis, MO), or a 1:200 dilution of mouse anti-syntaxin 6 antibody (Transduction Labs; Lexington, KY). Similarly, the secondary antibody also included a 1:200 dilution of fluorescein isothiocyanate (FITC)-conjugated donkey anti-mouse IgG (Jackson ImmunoResearch Laboratories).
Rat Islet Isolation and Culture
Pancreatic islets were isolated from 3–5-day-old Sprague-Dawley rats (Sasco; Omaha, NE) by a non-enzymatic culture method described previously (Hegre et al. 1983). 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). Before addition of hormones, the islets were cultured for 48 hr 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 on islet function was examined by culturing islets for 6 days 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 (BrdU) during the final 24 hr of culture (Brelje et al. 1993a). All islets were cultured at 37C in a humidified atmosphere of 5% CO2 in air.
Immunohistochemistry for STAT5 in Rat Islets
The localization of STAT5A and STAT5B in isolated rat islets was performed 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 anti-insulin antibody, a 1:1600 dilution of mouse monoclonal anti-glucagon antibody (Sigma), or a 1:400 dilution of mouse monoclonal anti-somatostatin (Sigma). The secondary antibodies also included a 1:600 dilution of cyanine 5.18 (CY5)-conjugated donkey anti-guinea pig IgG (Jackson Immunoresearch Laboratories). Because of the large size of the intact islets, they were also incubated with the secondary antibodies overnight at 4C. Glass beads of 50–100 μm maximal 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) (Brelje et al. 1989,1993b) mounted on an Olympus AX70 microscope equipped for epifluorescence (Lake Success, NY). The FITC and CY3 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. Brelje (available from the author). 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 maximal diameter of the nucleus was determined and the borders of the cell and its nucleus 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, 10–20 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 (NA 1.4). A total of at least 100 cells from three separate fields of view were analyzed for each experimental condition. The volumes and average fluorescence intensities of each cell, its cytoplasm, and its nucleus were calculated by tracing the borders on each image that contained part of an individual cell. In most experiments, the simpler technique of sampling only the cytoplasm and nucleus on a single image passing through the maximal diameter of the nucleus was used because it gave similar nuclear-to-cytoplasmic intensity ratios.
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 (NA 0.95). The average intensity values for the cytoplasm and nucleus for individual cells could not be compared because of a gradual attenuation of fluorescence intensities with increasing depth within the islets (Brelje et al. 1989,1993a,b). The absence of a correlation between the nuclear-to-cytoplasmic intensity ratio and the depth at which an image was acquired within an islet (r 2<0.1) demonstrates that calculation of the nuclear-to-cytoplasmic intensity ratio compensates for this artifact. Typically, five islets were analyzed for each experimental condition and each series of images contained 100–150 β-cells, 40–60 α-cells, and 10–20 δ-cells.
Immunohistochemistry for the PRL Receptor
A rabbit polyclonal antibody was raised against amino acids 45–65 of the extracellular portion of the rat PRL receptor (a 21-mer corresponding to the sequence pgtdgglptnysltyskegek) conjugated to bovine thyroglobulin by glutaraldehyde. Primary immunization was 1 mg of antigen in Freund's complete adjuvant injected IM into female New Zealand White rabbits. The first booster immunization with 1 mg in Freund's incomplete adjuvant was at 3 weeks, followed by subsequent immunizations at 2-week intervals. Serum was extracted from the immunized rabbits 10 days after each booster injection to monitor antibody titer. The final titer of the unfractionated serum was high enough for a 1:2000 dilution to detect bands of the appropriate molecular weight for the long and short forms of the PRL receptor (i.e., 85 and 49 kD) in Western blots prepared from rat islets. In addition, a 1:600 dilution was used to detect the staining of islets in cryostat sections prepared from adult rat pancreases. All staining was abolished when the antiserum was preadsorbed with the immunizing peptide.
Western Blotting Analysis
The Ins-1 cells (approximately 5 × 106) were cultured in 100-mm tissue culture dishes for 24–48 hr in complete medium. Before an experiment, the cells were cultured for an additional 24 hr in a defined serum-free medium and were at 50–60% confluence (see above). 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% NP-40 nonidet, 0.5% deoxycholate, 0.1% sodium dodecylsulfate, and 25 mM sodium fluoride supplemented with “Complete” protease inhibitor cocktail (Boehringer Mannheim; Indianapolis, IN) for 1 hr on a rocking bench at 4C. 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-ul 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 hr at 4C. 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 hr at 4C. 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% bovine serum albumin, 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 90C. 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 ensure protein transfer and to determine the location of molecular weight standards. The blots were incubated with 1% bovine serum albumin for 4–18 hr at 4C to block nonspecific staining, then with a 1:2000 dilution of the primary (i.e., probing) antibody overnight at 4C, and finally with a 1:30,000 dilution of either alkaline phosphatase-conjugated donkey anti-rabbit or anti-mouse IgG (Jackson ImmunoResearch Laboratories) for an additional 45 min at RT. 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 re-probed using the immunoprecipitating antibody.
Data Analysis and Presentation of Results
All results are expressed as means ± standard deviation (SD) of n observations. Statistical differences betweens the means were assessed by either Student's t-test or Student Neuman-Keuls one-way ANOVA for multiple comparisons.
Results
Effect of PRL on the Subcellular Localization of STAT5 in Ins-1 Cells
The effect of PRL on the subcellular localization of STAT5 was examined in Ins-1 cells exposed to 200 ng/ml PRL for 0 and 30 min. STAT5A and STAT5B localization was then determined by indirect IHC and acquiring a series of optical sections through the monolayers using a laser scanning confocal microscope. In unstimulated cells, both were detected throughout the cell as a fine punctate pattern with a slightly higher intensity in the nucleus compared to the cytoplasm (Figure 1, left). Prominent STAT5A immunoreactivity was also observed along the border of cells and in the perinuclear region of most cells. Although occasional cells had STAT5B immunoreactivity along their borders (Stout et al. 1997), an increase in STAT5B immunoreactivity was never observed in the perinuclear region. After PRL stimulation there was a marked increase in STAT5 immunoreactivity within the nucleus, with the relative increase for STAT5B being noticeably greater than for STAT5A (Figure 1, right). A decrease in the STAT5A immunoreactivity was also observed along the border of cells, with an increase in the perinuclear region. Curiously, the extent to which these non-nuclear changes in STAT5A distribution occurred with PRL stimulation was highly variable among different lots of this polyclonal antibody (see Discussion). In Ins-1 cells double-labeled with antibodies to various intracellular compartments, the STAT5A immunoreactivity in the perinuclear region overlapped with that for the transferrin receptor, a marker for early endosomal membranes, and with those for β-COP and syntaxin 6, markers for membranes of the trans-Golgi network (Figure 2). Although these observations suggest that PRL induced a redistribution of both STAT5s from the cytoplasm to the nucleus, other explanations are possible because a marked loss of immunoreactivity without a commensurate degradation of the respective STAT proteins after their activation has been reported (Rayanade et al. 1997,1998).
We then quantified the subcellular distribution of STAT5s to verify their nuclear translocation. For each cell, its entire volume was sampled by tracing the borders of the cell and its nucleus on each of the optical sections that contained part of the cell. This segmentation allowed the volume and total fluorescence intensity of individual cells, their nuclei, and their cytoplasm to be calculated (Table 1). This analysis demonstrated that the total cell fluorescence intensity for both STAT5s was unchanged by PRL stimulation (Figures 3A and B). This observation suggests that the ability of either antibody to detect its respective protein within the cells was not altered by activation of the STAT5s. Although heterogeneity in the response of individual cells was observed, the distribution profiles for the average intensity of the cytoplasm and nucleus suggest that all the cells responded (Figures 3C-3F). More importantly, the decrease in the total fluorescence intensity of the cytoplasm was identical to the increase in the nucleus after PRL stimulation for both STAT5s (Table 1). From these values, it was possible to estimate the proportion of each STAT5 that was translocated from the cytoplasm to the nucleus. PRL stimulation resulted in 12% of the STAT5A in the cytoplasm being translocated to the nucleus [i.e., (68–60)/68 = 0.12]). This represents a 25% increase in the amount of STAT5A in the nucleus [i.e., (40–32)/32 = 0.25]. Similarly, PRL stimulation resulted in 22% of the STAT5B in the cytoplasm being translocated to the nucleus [i.e., (60–47)/60 = 0.22]. This represents a 33% increase in the amount of STAT5B in the nucleus [i.e., (53–40)/40 = 0.33]. Therefore, the amount of the STAT5 was that translocated from the cytoplasm to the nucleus after PRL stimulation is surprisingly small, and most of the STAT5 present in the nucleus after PRL stimulation is still present in an inactive form.
Subcellular localization of STAT5 in Ins-1 cells after exposure to 200 ng/ml PRL for 30 min. The cells were stained with antibodies to either STAT5A (top) or STAT5B (bottom). PRL stimulated an increase in the staining intensity of the nucleus for both antibodies, with the relative difference being greater for STAT5B compared to STAT5A. Bar = 10 μm.
Although the relative changes in the intensity of the nucleus and/or cytoplasm can be used to monitor the nuclear translocation of STAT5, the ratio of the average pixel intensities of the nucleus and the cytoplasm was found to be a more sensitive indicator (Figure 4). This observation is a consequence of the volume of the nucleus being only one third the volume of the cytoplasm in the Ins-1 cells (Table 1). After recognizing the utility of the nuclear-to-cytoplasmic intensity ratio, we examined whether the ratio could be estimated from the average fluorescence intensities of the cytoplasm and nucleus on the single image passing through the maximal diameter of the nucleus. In unstimulated cells, the average nuclear-to-cytoplasmic intensity ratio for STAT5A was 1.00 ± 0.15 (n=120 cells) and for STAT5B was 1.23 ± 0.11 (n = 131 cells). After exposure to PRL for 30 min, the nuclear-to-cytoplasmic intensity ratio for STAT5A was slightly increased to 1.32 ± 0.26 (n=122 cells) and that for STAT5B was doubled to 2.44 ± 0.48 (n=137 cells). Because these values are identical to those previously calculated by sampling the entire volume of each cell (Table 1), this simpler technique was used for most experiments.
Effect of Antibody Concentration on the Nuclear-to-cytoplasmic Intensity Ratios
Because preliminary experiments suggested that the extent of nuclear translocation of STAT5 was highly variable among different lots of these polyclonal anti-bodies, we examined the effect of the primary antibody concentration on the nuclear-to-cytoplasmic intensity ratios observed in Ins-1 cells (Table 2). Although the nuclear-to-cytoplasmic intensity ratio was unchanged for most dilutions of each antibody in unstimulated cells, an increase in the nuclear-to-cytoplasmic intensity ratio was observed with lower concentrations of each antibody with PRL-stimulated cells. Therefore, the optimal dilution of each antibody was chosen to give the highest nuclear-to-cytoplasmic intensity ratio in PRL-stimulated cells without a substantial decrease in the overall intensity of staining. This experiment was repeated for each lot of these antibodies because a fourfold difference in the optimal dilution was observed. However, the maximal value for the nuclear-to-cytoplasmic intensity ratio after PRL stimulation was similar for different lots of the same antibody when used at its optimal dilution.
The subcellular localization of STAT5A in Ins-1 cells after exposure to 200 ng/ml PRL for 30 min. The cells were double-labeled with antibodies to STAT5A (red) and membrane markers for various intracellular compartments (green). The staining for the nuclear transport factor p97 (NTP, nuclear pores) shows that the STAT5A immunoreactivity is dispersed throughout the nucleus except for the nucleolus. The perinuclear staining for the transferrin receptor (TfR, endosomes), β-COP (trans-Golgi network), and syntaxin 6 (Syn6, trans-Golgi network) coincides with that observed for STAT5A (left). Bars = 5 μm. However, the punctate patterns for these markers within this region do not coincide with that observed for STAT5A when examined at a higher magnification (right). Bars = 2 μm.
Prolactin-stimulated translocation of STAT5 in Ins-1 cells
p<0.05 compared to the value for the corresponding control cells. Each value represents the mean ± SD.
Quantitative evaluation of the distribution of STAT5 in Ins-1 cells after exposure to 200 ng/ml PRL for 30 min. The total fluorescence intensity of individual cells (
Characterization of the STAT5 Antibodies by Western Blotting
Because of the high degree of structural homogeneity between STAT5A and STAT5B (Grimley et al. 1999), it was important to determine whether our observations were complicated by crossreactivity of the STAT5 antibodies. Their specificity was investigated in whole-cell lysates of Ins-1 cells that were immunoprecipitated and then blotted with each of the antibodies. The presence of homo- or heterodimmers of STAT5 in the lysates was precluded by using an extraction buffer containing both strong detergents and denaturants. The STAT5A immunoprecipitate contained a single band of similar molecular weight when probed with either the STAT5A antibody (Figure 5, Lane 1, top) or the STAT5B antibody (Figure 5, Lane 1, bottom). Because the molecular weight of STAT5B is slightly lower than that of STAT5A, this suggests that the probing STAT5B antibody is capable of recognizing STAT5A in Western blots. Therefore, it was not surprising that the STAT5B immunoprecipitate contained a single band of the molecular weight for STAT5A when probed with the STAT5A antibody (Figure 5, Lane 2, top) and additional bands of lower molecular weights when probed with the STAT5B antibody (Figure 5, Lane 2, bottom). This suggests that the STAT5B antibody is also capable of recognizing STAT5A when used for immunoprecipitations. In summary, these experiments indicate that the STAT5A antibody is highly specific for STAT5A and that the STAT5B antibody is also capable of recognizing STAT5A. This suggests that the nuclear translocation of STAT5B after PRL stimulation may be underestimated due to the crossreactivity of the STAT5B antibody with the more poorly translocated STAT5A.
The nuclear-to-cytoplasmic intensity ratios for STAT5 in Ins-1 cells after exposure to 200 ng/ml PRL for 30 min. The distributions for STAT5A (left) and STAT5B (middle) were calculated from the data presented in Figure 2. After PRL stimulation, the ratio was increased for both STAT5A and STAT5B (p<0.001; right). Each distribution represents more than 100 cells.
Effect of primary antibody concentration on the nuclear-to-cytoplasmic intensity ratio
p<0.001 compared to the value for the corresponding cells stained with 1.0 μg/ml for each antibody. Ins-1 cells were stained with either antibody after exposure to 200 ng/ml PRL for 30 min. Each value represents the mean ± SD for more than 100 cells.
Time Dependence of STAT5 Activation in Ins-1 Cells
The time dependence and correlation between the nuclear translocation and tyrosine phosphorylation of STAT5 in Ins-1 cells was also examined. The nuclear-to-cytoplasmic intensity ratios were determined for Ins-1 cells after incubation in the presence of 200 ng/ml PRL for 0–6 hr. An increase in the nuclear translocation of both STAT5A and STAT5B could be detected within 10 min and reached a maximum within 30 min (Figure 6, left). After this initial peak the nuclear translocation of STAT5B remained elevated, whereas the nuclear translocation of STAT5A was reduced to a level slightly above that observed before exposure to PRL (Figure 6, left). In parallel experiments it was not possible to detect the tyrosine phosphorylation of either STAT5A or STAT5B in Ins-1 cells before exposure to PRL (Figure 7). After PRL stimulation, the tyrosine phosphorylation of both STAT5A and STAT5B was biphasic, with an initial peak within 30 min, a nadir between 1–3 hr, and another increase after 4 hr that reached levels higher than those observed during the first peak (Figure 7). This pattern was not the result of changes in the amount of either STAT5 because their protein levels were unchanged (data not shown). In addition, the increases in the tyrosine phosphorylation of STAT5A were always observed slightly before those of STAT5B for both phases.
Characterization of the STAT5 antibodies by Western blotting. Lysates from Ins-1 cells were immunoprecipitated by either STAT5A (Lane 1) or STAT5B (Lane 2) antibodies. When the immunoprecipitate from STAT5A was probed by the STAT5A antibody, a single band corresponding to STAT5A was detected (Lane 1, top). This band was also detected by the STAT5B antibody (Lane 1, bottom). When the immunoprecipitate from STAT5B was probed by the STAT5A antibody, a single band corresponding to STAT5A was also detected (Lane 2, top). Multiple bands corresponding to STAT5B (lower bands, Lane 2, bottom) in addition to the slower migrating STAT5A (top band, Lane 2, bottom) were detected when probed with the STAT5B antibody.
Time dependence for the nuclear translocation of STAT5 after PRL stimulation in β-cells. Ins-1 cells were exposed to 200 ng/ml PRL for the indicated times and then stained with antibodies to either STAT5A or STAT5B (left). The nuclear-to-cytoplasmic intensity ratio was higher than controls for all times longer than 15 min (p<0.01). In rat islets exposed to 500 ng/ml PRL (middle), an increase in the nuclear-to-cytoplasmic intensity ratios was observed for all times after the addition of PRL (p<0.01). The reversal of the nuclear translocation of STAT5B was examined by exposing rat islets to 500 ng/ml PRL for 30 min and then incubating for the indicated times in the absence of PRL (right). The nuclear-to-cytoplasmic intensity ratio was higher than the controls at 0 and 15 min after the removal of PRL (p<0.01). Each point represents the mean ± SD of the nuclear-to-cytoplasmic intensity ratio from more than 100 cells for the Ins-1 cells or five islets.
Because JAK2 undergoes autophosphorylation on tyrosines after activation of the PRL receptor (Bole- Feysot et al. 1998), we also examined whether a similar pattern of changes in JAK2 activity occurs. It was not possible to detect the tyrosine phosphorylation of JAK2 in unstimulated cells (Figure 7). After exposure to PRL, the tyrosine phosphorylation of JAK2 was also biphasic, with an initial peak within 15 min, a nadir between 1 and 3 hr, and another increase after 4 hr that reached levels higher than those observed during the first peak (Figure 7). These observations suggest that changes in the biphasic pattern of STAT5 activation in Ins-1 cells are due to changes in JAK2 activity. Curiously, there was a marked decline in the protein levels of JAK2 after stimulation with PRL for 1 hr.
Effect of PRL on the Subcellular Localization of STAT5 in Rat Islets
To examine whether this approach for studying the activation of STAT5 could also be applied to normal β-cells within intact islets, isolated rat islets were exposed to 500 ng/ml PRL for 0 and 30 min, double-labeled with antibodies to either STAT5A or STAT5B and insulin to identify β-cells, and a series of optical sections acquired through the top of individual islets. In unstimulated islets, both were detected in all cells as a fine punctate pattern with a slightly higher intensity in the nucleus compared to the cytoplasm (Figure 8, left). The sequestration of STAT5A immunoreactivity in the perinuclear region was considerably less than that observed in the Ins-1 cells. Unexpectedly, the cytoplasm of some non-insulin-containing islet cells often had broad processes containing prominent STAT5A immunoreactivity (Figure 8, top). These cells were shown by double-labeling with somatostatin to be a subpopulation of the δ-cells (112 of the 176 δ-cells examined). After exposure to PRL there was an increase in STAT5 immunoreactivity within the nucleus, with the relative increase for STAT5B being noticeably greater than for STAT5A (Figure 8, right).
Time dependence of the tyrosine phosphorylation of JAK2, STAT5A, and STAT5B after PRL stimulation in Ins-1 cells. The cells were exposed to 200 ng/ml PRL for the indicated times and their lysates were immunoblotted and probed with antibodies to JAK2, STAT5A, STAT5B, and phosphotyrosine (PY; left). The blots were quantified for JAK2 phosphotyrosine and the amount of JAK2 protein (upper right), and STAT5A and STAT5B phosphotyrosine (lower right). The amount of STAT5A or STAT5B protein did not change during the experiment (data not shown).
We quantified the subcellular distribution of STAT5 by tracing the borders of each cell and its nucleus on the image passing through the maximal diameter of its nucleus. The average fluorescence intensity of these regions was then used to calculate a nuclear-to-cytoplasmic intensity ratio for each cell (Figure 9). Similar to the Ins-1 cells, an increase in the nuclear-to-cytoplasmic intensity ratios for both STAT5A and STAT5B was observed in β-cells (Table 3). Curiously, the distribution profile for STAT5B was much broader than that observed for STAT5A, suggesting a greater heterogeneity in the extent of response to PRL stimulation. These observations suggest that PRL also induces the nuclear translocation of both STAT5A and STAT5B in normal β-cells. It was necessary to estimate the amount of each STAT5 that was translocated from the cytoplasm to the nucleus after PRL stimulation in the β-cells of these isolated islets. The absolute amount could not be determined because the intensity values from the multiple images containing an individual cell could not be combined due to the gradual attenuation of the fluorescence intensities with increasing depth within the islets (Brelje et al. 1989,1993b). However, it was possible to estimate the proportion of each STAT5 that was translocated by estimating the volume of each cell and its nucleus from the size of their two-dimensional cross-sections used to calculate the nuclear-to-cytoplasmic intensity ratio. From these values, it appears that the volume of the cytoplasm was approximately five times the volume of the nucleus for normal β-cells (Table 3) compared to only twice as large for the Ins-1 cells (Table 1). Therefore, PRL stimulation was estimated to result in only 4% of the STAT5A in the cytoplasm being translocated to the nucleus [i.e., (77–74)/77 = 0.04]. This represents a 13% increase in the amount of STAT5A in the nucleus [i.e., (26–23)/23 = 0.13]. Similarly, PRL stimulation was estimated to result in 9% of the STAT5B in the cytoplasm being translocated to the nucleus [i.e., (77–70)/77 = 0.09]. This represents a 30% increase in the amount of STAT5B in the nucleus [i.e., (30–23)/23 = 0.30]. The translocation of only a small amount of each STAT5 in the cytoplasm was not unexpected because of the lower nuclear-to-cytoplasmic intensity ratios and larger disparity between the volume of the cytoplasm and nucleus of normal β-cells.
Subcellular localization of STAT5 in rat islets after exposure to 500 ng/ml PRL for 30 min. The islets were then stained with rabbit polyclonal antibodies to either STAT5A or STAT5B. The intensity of staining for the nucleus was increased after exposure to PRL in most of the cells within the islet. However, the relative difference was greater for STAT5B compared to the STAT5A antibodies. The cells containing prominent STAT5A immunoreactivity are δ-cells. Bar = 25 μm.
STAT5B Translocation in β-, α-, and δ-Cells
The absence of a nuclear translocation of STAT5B in the non-insulin-containing cells located at the periphery of the islets suggests that the other cell types are not PRL-responsive. To further examine this possibility, the nuclear-to-cytoplasmic intensity ratio was determined in rat islets exposed to 500 ng/ml for 30 min and double-labeled for STAT5B and either insulin, glucagon, or somatostatin (Figure 10). The nuclear-to-cytoplasmic intensity ratio was increased by PRL only in the insulin-containing β-cells, but not in the glucagon-containing α-cells or the somatostatin-containing δ-cells (Figure 11). In contrast, the nuclear-to-cytoplasmic intensity ratio was increased in all islet cell types when exposed to 10% fetal bovine serum for 30 min (Figure 11). This suggests that all islet cells have a functional STAT5 signaling pathway but that only β-cells are PRL-responsive. It is unclear which factors in fetal bovine serum stimulate the translocation of STAT5B in β-cells because it contains low amounts of PRL (most lots contain less than 15 ng/ml PRL in undiluted serum; HyClone, personal communication).
The PRL Receptor Is Expressed by β-Cells in Adult Rat Pancreas
Because the nuclear translocation of STAT5B was observed in β-cells only after PRL stimulation, it was of interest to determine whether the expression of the PRL receptor within islets was also limited to β-cells. A rabbit polyclonal antibody was prepared against a peptide corresponding to a 21-amino-acid sequence located in the extracellular domain of the rat PRL receptor as described in Materials and Methods. When this antibody was used to double-label tissue sections prepared from an adult rat pancreas, PRL receptor immunoreactivity was detected only in the insulin-containing β-cells but not in the glucagon-containing α-cells and the somatostatin-containing δ-cells (Figure 12). This restriction of the PRL receptor to β-cells is consistent with the ability of PRL to stimulate the nuclear translocation of STAT5B only in β-cells.
Quantitative evaluation of the distribution of STAT5 in rat islets after exposure to 500 ng/ml PRL for 30 min. The islets were stained with rabbit polyclonal antibodies to either STAT5A or STAT5B and the nuclear-to-cytoplasmic intensity ratios determined for individual cells. The distribution profiles for both STAT5A (left) and STAT5B (middle) are increased after PRL stimulation. The average nuclear-to-cytoplasmic intensity ratio was increased for each after PRL stimulation (p<0.001; right). Bars represent the mean ± SD of the nuclear-to-cytoplasmic intensity ratio from five islets.
Time Dependence of STAT5B Translocation in Rat Islets
To further characterize the nuclear translocation of STAT5B in normal β-cells, the time dependence for the translocation of STAT5B from the cytoplasm to the nucleus in response to PRL stimulation was examined. The nuclear-to-cytoplasmic intensity ratio was determined for β-cells in rat islets after incubation in the presence of 500 ng/ml PRL for 0–6 hr. An increase in STAT5B translocation could be detected within 5 min and reached a maximum within 30 min (Figure 6, middle). This initial peak was followed by a decrease in STAT5B translocation that reached a nadir between 1 and 3 hr. Subsequently, another increase in STAT5B translocation occurred after 3 hr, with a nuclear-to-cytoplasmic intensity ratio slightly less than that observed during the initial peak (Figure 6, middle). These experiments have been continued for up to 8 days without further changes in STAT5B translocation (data not shown). This biphasic pattern of STAT5B translocation in normal β-cells was much more prominent than that observed in the Ins-1 cells.
In similar experiments, the time dependence for the translocation of STAT5B from the nucleus to the cytoplasm after the removal of PRL was also examined. The nuclear-to-cytoplasmic intensity ratio was determined in β-cells in rat islets after incubation in the presence of 500 ng/ml PRL for 30 min and an additional 15, 30, and 60 min after the removal of PRL (Figure 6, right). A decrease in the nuclear-to-cytoplasmic intensity ratio was detected within 15 min and was indistinguishable from that observed in control islets by 60 min. The half-life for the decay of STAT5B translocation was estimated to be 27 min. This decay rate was independent of the length of the prior incubation in the presence of PRL for up to 48 hr (data not shown).
PRL-stimulated translocation of STAT5 in rat islets of Langerhans
p<0.05 compared to the value for the corresponding control cells. The percent of total cell content in the nucleus and the cytoplasm can be calculated from the corresponding nuclear-to-cytoplasmic intensity ratio and the estimated nucleus/cell volume ratio. An average nucleus/cell volume ratio of 0.22 ± 0.05 (n=20) was calculated using the average radius of 3.3 ± 0.2 μm for the nucleus and 5.5 ± 0.4 μm for the cytoplasm determined from the two-dimensional cross-sections and assuming that the three-dimensional volumes are spherical with the same radii. Each value represents the mean ± SD from five islets.
PRL stimulates the nuclear translocation of STAT5B only in the β-cells of rat islets. Isolated islets were exposed to 500 ng/ml PRL for 0 (top) and 30 min (middle and bottom) and triple-labeled with antibodies to STAT5B (red), insulin (green), and glucagon (blue; top and middle) or somatostatin (blue; bottom). The images on the left are projections across the top ~30 μm of an intact islet, whereas those on the right are single optical sections taken at a higher magnification from the same islets. Bars = 10 μm.
Dose Response for STAT5B Translocation in Rat Islets
If the nuclear translocation of STAT5B is an important event in the upregulation of islet function in response to PRL stimulation, the dose response for this event would be expected to be similar to those previously reported for the well-documented changes in islet function (Brelje and Sorenson 1997; Sorenson and Brelje 1997,2001). The nuclear-to-cytoplasmic intensity ratio was determined for β-cells in rat islets after incubation in the presence of 0–5000 ng/ml PRL for 30 min (Figure 13, left). The translocation of STAT5B could be detected with concentrations greater than 50 ng/ml PRL, with a maximal effect observed at 1000 ng/ml PRL. The ED50 was 130 ng/ml PRL. This dose response is remarkably similar to that observed for the increased insulin secretion and β-cell proliferation with rat islets cultured with PRL for 6 days (Figure 13, right). This suggests that the activation of STAT5 may be the key event in the upregulation of islet function in response to PRL stimulation.
Effect of PRL and fetal bovine serum on the subcellular localization of STAT5B in the β-, α-, and δ-cells of rat islets. Isolated islets exposed to 500 ng/ml PRL or 10% fetal bovine serum for 30 min and double-labeled with antibodies to STAT5B and either insulin, glucagon, or somatostatin to identify the different islet cell types. PRL stimulated the nuclear translocation of STAT5B only in the insulin-containing β-cells (p<0.01). In contrast, the fetal bovine serum stimulated the nuclear translocation of STAT5B in all islet cells (p<0.01). Bars represent the mean ± SD of the nuclear-to-cytoplasmic intensity ratio from five islets.
Discussion
The present study demonstrates that the activation of STAT5 can be monitored using an IHC method that examines its subcellular distribution within individual cells, in particular, that PRL stimulates the nuclear translocation of both STAT5A and STAT5B in the β-cells of normal rat islets of Langerhans. The nuclear-to-cytoplasmic intensity ratio was found to be particularly sensitive to small changes in its subcellular distribution. This occurs because of the reciprocal changes in STAT5 in the nucleus and the cytoplasm and because the volume of the nucleus is typically much smaller than the cytoplasm. For example, we were able to detect differences between groups with changes as small as 0.10–0.15 in the nuclear-to-cytoplasmic intensity ratio. This change corresponds to a redistribution of approximately 2% of the total cell content of STAT5. Because it is a ratio of two locally derived values, it also has the advantage of being less sensitive to variations in staining intensity and imaging artifacts across a specimen. The most important advantages of this approach are that individual cell types in a mixed population of cells can be examined and that it can be used with tissue sections to follow the activation of STAT5 in vivo.
The PRL receptor is expressed only in the β-cells of islets in the adult rat pancreas. Rat pancreas sections were double-labeled with antibodies to the PRL receptor (red) and one of the following islet hormones (green): insulin, glucagon, or somatostatin. Each image is a projection from 14–μm-thick sections. Bars = 50 μm.
A major limitation of using this IHC approach is the availability of high-quality antibodies. Preliminary studies demonstrated that relatively few of the commercially available antibodies were suitable for showing the nuclear translocation of STAT5 even though the same antibodies could detect STAT5 on Western blots. This difficulty was most likely related to the low titer and/or the high background staining with many of these antibodies. In general, affinity-purified antibodies were superior. These experiments were further complicated because STAT5A and STAT5B are closely related proteins that share 96% homology (Azam et al. 1995; Liu et al. 1995; Mui et al. 1995). These proteins differ at their COOH terminus, a highly variable region among other STAT proteins that appears to be involved in transcriptional activation. In Western blots, it was shown that the STAT5A antibody was highly specific for STAT5A, whereas STAT5B antibody showed evidence of crossreactivity with STAT5A. It is unknown whether a similar extent of crossreactivity occurs in fixed tissue. However, the impact of this crossreactivity would be diminished if the relative abundance of STAT5B is higher than STAT5A, as has been suggested for Ins-1 cells (Galsgaard et al. 1999) and shown for the liver (Ripperger et al. 1995; Ram et al. 1996; Park et al. 1999). The failure to observe an increase in STAT5B immunoreactivity in the perinuclear region of Ins-1 cells, which contains the most intense STAT5A immunoreactivity, suggests that this crossreactivity is less pronounced in fixed tissue or that the relative abundance of STAT5B is much higher than STAT5A in islet cells. More troubling was the high variability among different lots of the STAT5A antibody in showing changes in STAT5A immunoreactivity after PRL stimulation. Because each lot of this polyclonal antibody is pooled from multiple rabbits that are continually replaced (Santa Cruz Biotechnology, personal communication), this variability suggests that only a subset of the epitopes to the carboxyl terminus of STAT5A may be involved in these non-nuclear changes in STAT5A distribution.
Dose response for the nuclear translocation of STAT5B after PRL stimulation in rat islets. The nuclear-to-cytoplasmic intensity ratio was determined in islets after exposure to between 0 and 5000 ng/ml PRL for 30 min (left). This dose response is similar to that observed for the effects of PRL for 6 days on insulin secretion and islet cell proliferation (right). An increase in the nuclear-to-cytoplasmic intensity ratio, insulin secretion, and islet cell proliferation was observed for all PRL concentrations of 100 ng/ml or higher (p<0.01).
The expression of both STAT5A and STAT5B by most cells suggests that each may have specific functions (Grimley et al. 1999). Although both STAT5s bind to canonical GAS sites with the general structure TTCNNNNGAA (Ihle 1996), differences in the DNA-binding specificities of their homodimers have been demonstrated (Boucheron et al. 1998; Verdier et al. 1998). This similarity in structure and function made the study of mice with targeted STAT5 gene disruptions particularly interesting. PRL-directed mammary gland maturation fails without functional STAT5A (Liu et al. 1997; Teglund et al. 1998), while disruption of STAT5B in males abrogates the effects of pulsatile growth hormone on hepatic function and body mass (Udy et al. 1997; Park et al. 1999). Although these defects can be ascribed to non-redundant functions of each STAT5, it is also possible that some of these phenotypic differences may be related to their expression levels in different tissues (Teglund et al. 1998). For example, the mammary glands of STAT5A-deficient mice can be partially restored to function through compensatory hyperactivation of STAT5B after repeated cycles of pregnancy (Liu et al. 1998). In addition, the effect on hepatic function in STAT5B-deficient mice is undoubtedly related to the much higher relative abundance of STAT5B compared to STAT5A in the liver (Ripperger et al. 1995; Ram et al. 1996; Park et al. 1999). Therefore, the specific function and functional redundancy of these two forms of STAT5 has not yet been determined.
Interestingly, differences in the subcellular localization of STAT5A and STAT5B were observed in Ins-1 cells. STAT5A immunoreactivity was detected along the border of unstimulated cells and in the perinuclear region of both unstimulated and stimulated cells. Similarly, caveolae found in the plasma membrane of endothelial cells have been shown to contain the bradykinin 2 receptor, the Tyk2 kinase, and STAT3 before stimulation with bradykinin (Ju et al. 2000); and an increase in STAT1 and STAT2 within the Golgi complex of the intestinal cell line IEC-6 has been reported after stimulation with epidermal growth factor (Johnson et al. 1999). These observations suggest that STAT5A is associated with the PRL receptor on the surface of unstimulated cells and then follows the internalization of the ligand-bound PRL receptors after stimulation. It is unclear why a similar redistribution of STAT5B in the cytoplasm of the Ins-1 cells was not observed.
There also appeared to be a difference in the amount of STAT5A compared to STAT5B translocated to the nucleus after PRL stimulation. For Ins-1 cells, 12% of the STAT5A and 22% of the STAT5B located in the cytoplasm was translocated to the nucleus. For the β-cells in rat islets, it was estimated that 4% of STAT5A and 9% of STAT5B located in the cytoplasm was translocated to the nucleus. Curiously, in both cases the amount of STAT5A translocated from the cytoplasm to the nucleus was approximately half of the amount of STAT5B. This difference may be underestimated in our experiments due to the crossreactivity of the STAT5B antibody with the more poorly translocated STAT5A. A similar relationship between the nuclear translocation of STAT5A compared to STAT5B was also observed with rabbit polyclonal antibodies from Upstate Biotechnology (unpublished observations). Because the calculation of the nuclear-to-cytoplasmic intensity ratio should be insensitive to differences in the relative abundance of STAT5A and STAT5B in β-cells, the reasons for this imbalance are unclear. The high degree of structural similarity between the SH2-binding domains of STAT5A and STAT5B suggests that differences in the binding affinities for the tyrosine-phosphorylated docking sites on the PRL receptor are not responsible (Grimley et al. 1999). In addition, the nuclear translocation of such a small amount of each STAT5 also suggests that the pool of cytoplasmic STATs available for tyrosine phosphorylation remains quite large and was not rate-limiting even after prolonged stimulation.
Another difference between the activation of STAT5A and that of STAT5B was apparent when the time dependence of their tyrosine phosphorylation and nuclear translocation was examined in β-cells. The nuclear translocation of STAT5B in normal β-cells after PRL stimulation was biphasic, with an initial peak within 30 min and prolonged activation after 4 hr. Although less pronounced, this pattern was also observed in Ins-1 cells. In contrast, the nuclear translocation of STAT5A peaked within 30 min and was then reduced to a level slightly above that observed before the exposure to PRL in the Ins-1 cells. Curiously, the tyrosine phosphorylation of both STAT5A and STAT5B had similar biphasic profiles. The reason for this discrepancy between the tyrosine phosphorylation and nuclear translocation of STAT5A is unknown. It should also be noted that these biphasic changes in the activation of STAT5B are different from the transient activation of STATs observed with many cytokines (Ihle 1996; Darnell 1997), such as growth hormone (Ram et al. 1996; Gebert et al. 1997). Furthermore, these changes were correlated with similar changes in the activation of JAK2, as determined by its tyrosine phosphorylation.
This study also demonstrated that PRL stimulates the nuclear translocation of STAT5 only in the β-cells of normal rat islets. The presence of a functional STAT5 signaling pathway in the glucagon-containing α-cells and the stomatostatin-containing δ-cells was demonstrated by incubation of the islets with fetal bovine serum. This observation was unexpected because we have previously reported the presence of the PRL receptor on both α- and β-cells using mouse monoclonal antibodies to the PRL receptor (Sorenson and Stout 1995). To reexamine this issue, we generated a polyclonal rabbit antibody against amino acids 45–65 of the extracellular portion of the rat PRL receptor. Use of this antibody suggested that only β-cells, but not the α-cells or δ-cells, express PRL receptors. Apparently, the staining of the α-cells with the previous antibody was nonspecific and was most likely due to the high affinity of these cells for a number of antibodies (unpublished observation). It is interesting to note that this study also observed an increase in immunoreactivity for the PRL receptor in β-cells, but not in α-cells, in islets isolated from pregnant rats (Sorenson and Stout 1995). Because activation of the PRL receptor increases the expression of the PRL receptor by β-cells (Moldrup et al. 1993; Galsgaard et al. 1999), the failure to observe an increase in α-cells also suggests that this staining was nonspecific. Therefore, the present results are compatible with the changes in islet function during pregnancy being limited to those that occur in β-cells. It is also intriguing that the dose response for the nuclear translocation of STAT5B by PRL was remarkably similar to that observed for the increased insulin secretion and β-cell proliferation with rat islets cultured in the presence of PRL for 6 days. This observation suggests that the activation of STAT5 may be the key event in the upregulation of islet function in response to activation of the PRL receptor. Furthermore, the induction of a dominant-negative STAT5 mutant under the control of doxycycline-inducible promoter completely inhibited the effects of PRL on Ins-1 cells (Friedrichsen et al. 2001).
Footnotes
Acknowledgements
Supported by NIH grant DK33655.
We would like to thank Dr A.F. Parlow (National Hormone and Pituitary Program of the NIDDK) for providing the rat PRL used in this study.