Expression and Distribution of Vascular Endothelial Growth Factor Receptor Flk-1 in the Rat Pituitary

Abstract

Vascular endothelial growth factor (VEGF) acts primarily as an endothelial cell mitogen via the specific receptors Flk-1 and Flt-1. To help further define the possible role of VEGF in the control of pituitary cell function, we examined Flk-1 expression in normal rat pituitaries and in GH3 cells. Flk-1 expression was studied by immunohistochemistry, in situ hybridization, and double-labeling immunofluorescence combined with confocal laser microscopy. In normal rat pituitaries, Flk-1-immunoreactive cells appeared widely distributed only in the anterior lobe and were not detected in the intermediate or posterior lobe. Apart from the adenohypophysial cells, Flk-1 immunopositivity was also evident in endothelial cells of many capillaries distributed throughout the gland. Immunohistochemistry also showed that majority of GH3 cells expressed Flk-1 protein. In situ hybridization showed conclusive staining with the antisense probe and confirmed the immunohistochemical results. The double immunofluorescence method revealed Flk-1 expression in all types of hormone-producing adenohypophysial cells but not in folliculostellate cells. The percentage of immunopositive cells varied among the various cell types. The present study demonstrates that pituitary cells are not only sources of VEGF but also targets of this multifunctional substance, supporting the concept that VEGF functions as an autocrine/paracrine factor in the pituitary.

Vascular endothelial growth factor (VEGF) is a multifunctional polypeptide initially thought to be mainly a mitogen specific for endothelial cells (Ferrara et al. 1992; Ferrara and Davis-Smyth 1997). It has since been found to mediate a number effects on both endothelial and non-endothelial cells (Dvorak et al. 1995). For example, VEGF enhances vascular permeability and facilitates transient accumulation of cytoplasmic calcium in endothelial cells (Dvorak et al. 1995; Seymour et al. 1996). Previous studies also suggested the existence of a paracrine/autocrine loop involving VEGF in both proliferation and differentiation of various normal and neoplastic cells, including pancreatic duct epithelium and pancreatic cancerous cells (Oberg et al. 1994; Oberg-Welsh et al. 1997; Itakura et al. 2000).

Although the high-affinity tyrosine kinase receptors Flk-1 (fetal liver kinase-1), the equivalent of human KDR (kinase insert domain-containing receptor), and Flt-1 (fms-like tyrosine kinase) are specific targets of VEGF activity, it appears that both the in vivo and in vitro effects of VEGF are mediated almost exclusively via the Flk-1 receptor (Dvorak et al. 1995). The VEGF/Flk-1 receptor system plays an important role in the growth of pancreatic tumors, regulating cancer cell division in an autocrine/paracrine fashion (Von Marschall et al. 2000). In addition, Yoshiji et al. (1999) showed the Flk-1 receptor to be a major regulator in the development of VEGF-induced murine hepatocellular carcinoma.

In the rat, previous immunohistochemical studies have reported pituitary overexpression of both VEGF and Flk-1 receptor after 7 days of estrogen exposure. This observation suggests a regulatory role of the VEGF/Flk-1 system in estrogen-induced murine pituitary tumors (Banerjee et al. 1997). Although VEGF is widely distributed throughout the normal adenohypophysis and its tumors, in the rodent and human (Banerjee et al. 1997,2000; Lloyd et al. 1999; Ochoa et al. 2000; Vidal et al. 1999,2000a) previous studies demonstrated Flk-1 receptor expression only in endothelial cells of pituitary capillaries, thus suggesting that the effect of VEGF is limited to the pituitary (Banerjee et al. 1997).

To obtain deeper insight into the role of the VEGF/Flk-1 system in anterior pituitary cell activity, we investigated expression of Flk-1 receptor mRNA and of its protein in normal rat pituitaries and in GH3 cells, a rat pituitary tumor cell line secreting both prolactin and growth hormone. Immunohistochemistry combined with confocal laser microscopy was also used to identify those pituitary cells representing targets of VEGF activity.

Materials and Methods

Animals and Tissue Preparation

All rats were handled in accordance with the principles and procedures of the Guiding Principles for the Care and Use of Animal Research (86/609/EU). Six 3-month-old female Sprague-Dawley rats (220–250 g) in the proestrous phase of their cycle were housed at room temperature (RT) with a controlled photoperiod and were fed a diet of rat pellets and tapwater ad libitum. The stage of their cycle was determined by daily vaginal smears.

The animals were deeply anesthetized with pentobarbital and then perfused via the ascending aorta with 4% paraformaldehyde in PBS. Their pituitaries were immediately removed, bisected sagittally, and fixed by immersion in 4% paraformaldehyde in PBS at 4C for 2 hr. After fixation, half of each pituitary was dehydrated and embedded in paraffin using an automatic processor. The other half of the gland was immersed in 15% sucrose in 0.1 M phosphate buffer for at least 24 hr, then quickly frozen and cut into 10-μm-thick sections in a cryostat for co-localization studies of Flk-1, the VEGF receptor.

Cell Culture

Obtained from the American Type Culture Collection (ATCC; Rockville, MD), the GH3 cells were grown in 175-ml flasks containing DMEM supplemented with 10% (v/v) horse serum, 5% (v/v) fetal calf serum, 48 mg penicillin/liter, and 100 mg streptomycin/liter at 37C in an atmosphere of 5% CO₂ in air at 100% humidity. The GH3 cells were plated in culture wells at a density of 10–20 × 10⁴ cells per well. To facilitate cell attachment, they were maintained for 48 hr in supplemented DMEM. Before fixation, cells were washed twice with Ham-I solution (122 mM NaCl, 30 mM HEPES, 1 mM NaPO₄H₂·2H₂O, 3 mM KCl, pH 7.4). These conditions increased the uniformity of the results. Thereafter, cell aliquots were fixed in 4% paraformaldehyde for 30 min at RT and attached to poly-L-lysine coated slides.

Immunohistochemistry and In Situ Hybridization

IHC and ISH procedures were performed to assess localization of Flk-1 protein and its mRNA in serial sections of paraffin-embedded rat pituitaries and in the GH3 cell line.

For the ICC demonstration of the Flk-1 receptor, a monoclonal antibody directed against a peptide sequence of the carboxyl terminal portion of the mouse Flk-1 receptor (Santa Cruz Biotechnology, SC-6251; Santa Cruz, CA) was used. Preliminary titration experiments determined the optimal working dilution (1:100). To ensure specificity, control tests included both replacement of Flk-1 antibody with TRIS-buffered saline (TBS) and preabsorption of the Flk-1 antibody with homologous antigen (Santa Cruz Biotechnology SC-315P). Immunolabeling was completely abolished in both controls.

ICC studies utilized the streptavidin-biotin-peroxidase complex technique (Hsu et al. 1981) with slight modifications. Before application of the primary antibody, paraffin sections were pretreated with 1.5% H₂O₂ in absolute methanol for 30 min to inactivate endogenous peroxidase activity. Thereafter, 10% normal horse serum was applied for 1 hr to block nonspecific binding of the secondary antibody. Lastly, the tissue was incubated overnight with the monoclonal Flk-1 antibody at RT. Slides were then washed three times (5 min each) in TBS and incubated for 1 hr with biotinylated horse anti-mouse IgG (Vector BA-2000; Burlingame, CA) diluted 1:100. The buffer rinses were then repeated and the slides were incubated for 1 hr with the streptavidin-peroxidase complex (Sigma E-2886; St Louis, MO) diluted 1:100. The final reaction was achieved by incubating the sections for 5 min in a solution containing 5 mg DAB and 1 ml 1% H₂O₂ in 100 ml TRIS buffer (pH 7.6). The slides were then counterstained with hematoxylin, coverslipped, and examined.

ISH was performed using an oligonucleotide probe complementary to regions 309–338 of mFlk-1. The antisense stand sequence of the oligonucleotide is 5‘-CTC GTT CCA CGA CGA CCG GCA GCG GGA CAC-3’. This was synthesized by the solid-phase cyanoethyl phosphoramidiete method on an automated DNA synthesizer (Gene Assembler; Pharmacia, Hoefer Pharmacia Biotech, San Francisco, CA) and then purified by gel electrophoresis on 20% polyacrylamide gels. Paraffin sections were dewaxed in xylene, rehydrated, and transferred to diethylpyrocarbonate water (DEPC water). Then the sections were pretreated with proteinase K (1 μg/ml) at 37C for 15 min before prehybridization for 1 hr at RT in hybridization mixture not containing the probe. Overnight hybridization was then performed with the digoxigenin-labeled oligonucleotide probe at 37C. Hybridized sections were then washed three times (5 min each) with 2 × SSC, followed by 1 × SSC and finally by 0.5 × SSC. After a rinse in Buffer A (100 mM Tris-HCl, 150 mM NaCl, pH 7.5) the sections were bathed in blocking reagent (Buffer A with 1% normal sheep serum and 0.3% Triton X-100) for 30 min and incubated with anti-digoxigenin-alkaline phosphatase complex (1:200; Boehringer Mannheim 1093274, Mannheim, Germany) for 3 hr. Three washes in Buffer A and one in Buffer C (50 mM Tris-HCl, 100 mM NaCl, 50 mM MgCl₂, pH 9.5) were followed by incubation in a substrate solution containing 4-nitroblue tetrazolium chloride, 5-bromo-4-chloro-3 indolyl phosphate, and levamisole, all in Buffer C. The reaction was stopped in Buffer C. The slides were then counterstained with nuclear fast red, coverslipped, and examined.

A digoxigenin-labeled oligonucleotide sense probe (5‘-GAG-CAAGGTGCTGCTGGCCGTCGCCCTGTG-3’) served as a control to confirm the specificity of the ISH procedure.

Double-labeling Immunofluorescence and Confocal Laser Microscopy

Double-labeling immunofluorescence was applied to specifically identify the cell type(s) expressing Flk-1. Double immunostaining was performed on cryostat sections by combining monoclonal anti-Flk-1 antibody with a rabbit polyclonal antibody against pituitary hormones or S-100 protein, a marker of folliculostellate cells. Specifics of the various anti-sera employed were as follows: rabbit antisera directed against rat prolactin (PRL) (dilution 1:500; National Hormone and Pituitary Program, NIDDK-AFP-4251091, Baltimore, MD), rat growth hormone (GH) (dilution 1:500; UCB Bioproducts, i0538/002, Braine-L'Alleud, Belgium), human adrenocorticotropic hormone (ACTH) (dilution 1:400; Sigma A-0673), rat thyroid-stimulating hormone (β-TSH) (dilution 1:400; National Hormone and Pituitary Program, NIDDK-AFP-1274789), human luteinizing hormone (β-LH) (dilution 1:400; National Hormone and Pituitary Program, NIDDK-AFP-5595), human follicle-stimulating hormone (β-FSH) (dilution 1:400; National Hormone and Pituitary Program, NIDDK-AFP-20102688), and bovine S-100 protein (1:400, Sigma S-2644). After rinsing in TBS, the double-stained sections were incubated at RT for 1 hr with FITC-conjugated donkey anti-mouse (dilution 1:250; Jackson Immunoresearch Laboratories 715-095-151, West Grove, PA) and rhodamine red-X-conjugated donkey anti-rabbit (dilution 1:250, Jackson Immunoresearch Laboratories 711-085-152). After rinsing in TBS, the sections were mounted in a mixture of PBS and glycerol (1:3) containing 0.1% p-phenilenediamine to prevent fading of the immunofluorescence reaction. Sections were examined on an MRC 1024 ES con-focal laser scanning system (Bio-Rad Laboratories; Hercules, CA). The excitation wavelength was 488 nm for FITC and 514 nm for rhodamine red-X-induced fluorescence. Specificity studies carried out by omitting primary antisera or by preabsorbing the primary antisera with homologous antigen excess; all showed the absence of the fluorescent signal. As an additional control, we tested whether the secondary antisera used in the double labeling were specific and did or did not crossreact with each other. To test for this possibility, one of the primary antisera was incubated alone and revealed by both secondary antisera. The nonspecific staining was totally abolished without alteration of the specific staining.

Double immunofluorescence was used to determine the proportions of Flk-1-immunopositive cells, those immunoreactive for either adenohypophysial hormones or S-100 protein. In each section, 15 randomly selected fields were studied on a confocal laser scanning system at a magnification of ×60. Approximately 100 cells were assessed in each field; only nucleated cells were counted.

Data were tested for statistical significance using the SPSS Statistical Computer Program (SPSS; Chicago, IL). As a multiple comparison method, all data were evaluated by one-way ANOVA and the Student's t-test. Only differences of p<0.05 were considered statistically significant.

Results

Pituitary Glands

In the six rat pituitaries, Flk-1-immunoreactive cells appeared widely distributed only in the anterior lobe and were not detected in the intermediate or posterior lobes. Immunoreactivity was moderate to strong and involved the entire cytoplasm (Figure 1A and 1B). Apart from the adenohypophysial cells, Flk-1 immunopositivity was also evident in endothelial cells of many capillaries distributed throughout the gland (Figure 1B).

ISH showed conclusive staining with the antisense probe, whereas the sense probe preparation was negative. Hybridization labeling was in the same pattern as the IHC results. Thus, Flk-1 mRNA signal appeared widely expressed throughout the adenohypophysis. In positive cells, the signal was moderate to strong and cytoplasmic. The signal was also detected in endothelial cells of most capillaries throughout the gland (Figure 1C).

GH3 Cells

IHC showed that the majority of GH3 cells expressed Flk-1 protein. The pattern of immunopositivity was diffuse, moderate to strong cytoplasmic staining. ISH also demonstrated Flk-1 mRNA in the majority of GH3 cells. Strong cytoplasmic labeling was observed (Figure 1D).

Double-labeling Immunofluorescence and Confocal Laser Microscopy

Because the Flk-1 receptor was diffusely distributed in the normal adenohypophysis, double-labeling immunofluorescence combined with confocal laser microscopy was applied to identify which specific cells expressed the receptor (Figures 2 and 3). The double-immunofluorescence method demonstrated that all types of hormone-producing adenohypophysial cells were immunopositive for Flk-1. In contrast, the protein was not detected in folliculostellate cells. The percentage of immunopositive cells varied considerably among the various cell types (Figure 4). Co-localization of Flk-1 and pituitary hormones was particularly abundant in somatotrophs and lactotrophs (60-50%). Significantly fewer cells were immunopositive among corticotrophs (40-30%), thyrotrophs (10-5%), and gonadotrophs (10–5%) (Figure 4).

Discussion

In addition to adenohypophysial hormones, pituitary cells contain various other substances including many growth factors, cytokines, angiogenic substances, and neuropeptides. Frequent expression of the corresponding receptors on adenohypophysial cells enables these factors to influence cell growth and function by auto-or paracrine mechanisms (Ray and Melmed 1997; Spada 1998). An intrinsic intercellular communication network appears to be involved in the control of pituitary homeostasis, particularly with regard to gene expression and the production/secretion of anterior pituitary hormones (Arzt et al. 1998,1999). VEGF is also an autocrine/paracrine molecule secreted by normal and neoplastic anterior pituitary cells (Banerjee et al. 1997,2000; Lloyd et al. 1999; Vidal et al. 1999, 2000a; Ochoa et al. 2000). Although unanswered questions remain regarding the functional signifcant of VEGF expression in the adenohypophysis, it is conceivable that VEGF plays a role both in the formation of pituitary portal vessels during fetal life and in maintenance of their differentiated state in mature animals (Ferrara et al. 1992). Various lines of evidence indicate that VEGF is involved in pituitary neovascularization in estrogen-treated rats and in the course of development of prolactin cell adenomas (Banerjee et al. 1997,2000). It has also been suggested that VEGF affects the endocrine activity of pituitary cells. The latter is supported by the recent finding that VEGF is stored within secretory granules of the normal human pituitary cells, in which it is co-localized with the full spectrum of adenohypophysial hormones (Vidal et al. 1999). This subcellular distribution of VEGF indicates that it is simultaneously released with the various pituitary hormones and perhaps affects vascular permeability and the transport of adenohypophysial hormones across capillaries. The finding of VEGF expression in GH3 cells (Banerjee et al. 2000; Ochoa et al. 2000; Vidal et al. 2000a) also supports the notion that VEGF plays a specific role(s) in the control of endocrine secretion and proliferative activity of pituitary cells.

Figure 1

(A,B) Immunocytochemical staining for Flk-1 receptor in the pituitary of proestrous rat. (A) Immunopositive cells (arrowheads) are randomly distributed throughout the anterior lobe. (B) The endothelial cells of blood vessels are well labeled (arrowheads). (C,D) Localization of Flk-1 expression within the pituitary of proestrous rat shown by ISH. (C) Pituitary section was hybridized with digoxigenin-Flk-1 antisense probe. Labeling for Flk-1 is evident in the adenohypophysial cells (arrowheads) and endothelial cells (arrow). Asterisk, vascular lumen. (D) The sense probe is negative. (E,F) Localization of Flk-1 expression within GH3 cells shown by ISH. (E) Cytospin preparation of GH3 cells showing a strong cytoplasmic hybridization signal. (F) The sense probe is negative. Bars = 10 μm.

Our study demonstrates that pituitary cells are not only sources of VEGF but also targets of this multifunctional substance. It also supports the concept that this cytokine functions as an autocrine/paracrine factor in the pituitary. Previous studies have shown that certain angiogenic factors, including epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF), play a regulatory role in adenohypophysial hormone gene expression. Although on the basis of our data it is not possible to determine the biological significance of Flk-1 expression in pituitary cells, it appears that the Flk-1 receptor acts as a key modulator of VEGF activity not only in endothelial cells but in endocrine cells as well. Supporting this idea are previous studies showing that the VEGF/Flk-1 system plays a regulatory role in several physiological and pathological processes. In the fetal rat pancreas, Oberg-Welsh et al. (1997) demonstrated the Flk-1 receptor in the duct epithelium and islet-like structures, suggesting that the VEGF/Flk-1 system affects pancreatic morphogenesis by stimulating the development of endocrine precursor cells in pancreatic duct epithelium. It was also shown that VEGF and its receptor (Flk-1) are widely expressed in neoplastic pancreatic cells, suggesting that VEGF in an autocrine/paracrine fashion is implicated in the control of development and growth of pancreatic tumors (Christofori et al. 1995; Itakura et al. 2000; Von Marschall et al. 2000). The view that the VEGF/Flk-1 system plays an important role in the regulation of normal and neoplastic cell growth is in keeping with the signal transduction mechanism stimulating a protein kinase receptor. The Flk-1 receptor is tyrosine-phosphorylated when stimulated with its appropriate ligand (VEGF). Phosphorylated tyrosine residues control the kinase activity of the receptor, leading to an intracellular cascade and activation of the MAPK (mitogen-activated protein kinase) pathway resultant cell division (Ullrich and Schlessinger 1990; D'Angelo et al. 1995). On the basis of these findings, it is reasonable to assume that the VEGF/Flk-1 system can induce cell division in the pituitary. It should be noted, however, that the mitogenic response to VEGF is cell type-specific. Thus it was demonstrated that, in contrast to endothelial cells the effect of VEGF on fibroblast MAPK activation is delayed and that the mitogenic response is weaker (Takahashi and Shibuya 1997).

Figure 2

Double-label immunofluorescent staining combined with confocal laser microscopy in the pituitary of proestrous rat. White arrowheads indicate cells double labeled for Flk-1 (green) and the respective pituitary hormone, growth hormone (A-C), prolactin (D-F), and adrenocorticotropic hormone (G-I) (red). White arrows show adenohypophysial cells stained for any of the pituitary hormones but not for Flk-1. Note that the green Flk-1 label is not identified in the intermediate lobe (IL). Bars = 10 μm.

Figure 3

Expression of Flk-1 in immunocytochemically identified adenohypophysial cells in the pituitary of proestrous rat. White arrowheads indicate double-labeled cells with Flk-1 visualized in the green immunofluorescent channel and hormones (white arrows), thyroid-stimulating hormone-β (A-C) and luteinizing hormone-β (D-F), labeled in the red immunofluorescent channel. White arrows show adenohypophysial cells stained for any of the pituitary hormones or S-100 protein but not for Flk-1. Note that the green Flk-1 label is not detected in folliculostellate cells (G-I). Bars = 10 μm.

Figure 4

Percentage of hormone-producing adenohypophysial cells immunopositive for Flk-1. Data are expressed as mean + SEM (%). Values with no letters in common are significantly different, ^∗ p<0.05 (statistical analysis with one-way ANOVA and Student's t-test).

Our study demonstrates that in the normal rat adenohypophysis Flk-1 receptor is expressed in all but folliculostellate cells. The absence of Flk-1 receptor in pituitary folliculostellate cells is intriguing because this cell type is the main source of VEGF in the normal adenohypophysis (Gospodarowicz and Lau 1989; Ferrara and Davis-Smyth 1997; Jabbour et al. 1997). These observations provide the basis for a novel mechanism of VEGF cell-to-cell interaction between hormone-secreting and folliculostellate cells. Previous data regarding paracrine communication among pituitary secretory cells supported the role of local factors in the regulation of anterior pituitary hormone secretion. The pathway includes protein accumulation in secretory granules, their transportation to the cell periphery, and discharge of granule contents by either exocytosis or transmembrane diffusion. It is well known that stimulation of adenohypophysial cell secretion is accompanied by alterations in ion concentration that result in rapid, reversible changes in the membrane potential and subsequent resultant modulation of the cytoplasmic Ca²⁺ concentration (Kwiecien and Hammond 1998; Tse and Tse 1999). Because the VEGF/Flk-1 system is a key mediator of vascular permeability and cytosolic Ca²⁺ concentration in the endothelial cells (Dvorak et al. 1995; Seymour et al. 1996), specific expression of Flk-1 in hormone-producing adenohypophysial cells suggests the participation of the VEGF/Flk-1 system in the control of endothelial cell permeability and hormone secretion.

Our findings of Flk-1 receptor expression in GH3 cells implicates VEGF in the regulation of neoplastic anterior pituitary cell function. Banerjee et al. (2000) recently showed that estrogen-induced rat pituitary tumors and the GH3 pituitary cell line express VEGF164, an isoform belonging to the VEGF family, and its co-receptor, neuropilin-1. This suggests that VEGF may modulate the behavior of rat pituitary tumor cells. Although it is not yet known whether Flk-1 or other VEGF receptors are expressed in human adenohypophysial tumors, previous studies have demonstrated that VEGF is widely expressed in most human pituitary adenomas and carcinomas. Because pituitary tumors have limited angiogenic capacity (Turner et al. 2000a,b; Vidal et al. 2000b), it is conceivable that VEGF might act on pituitary tumor cells through an autocrine or paracrine mechanism. In accordance with this idea is our finding that VEGF expression is highest in GH adenomas, the least vascularized of adenohypophysial tumor types (Lloyd et al. 1999). We have also demonstrated that pituitary carcinomas show increased VEGF expression compared to adenomas. This suggests VEGF upregulation during the process of transformation of pituitary adenomas to carcinomas.

Clearly, further studies are required to clarify the functional role of the VEGF/Flk-1 system in the control secretory activity and growth in normal and neoplastic pituitary tissue. Such studies may bring about the use of angiogenic and anti-angiogenic factors in the control of medical and neoplastic diseases of the pituitary gland.