Vascular Endothelial Growth Factor Up-regulation via p21-activated Kinase-1 Signaling Regulates Heregulin-β1-mediated Angiogenesis*

Heregulin-β1 promotes the activation of p21-activated kinase 1 (Pak1) and the motility and invasiveness of breast cancer cells. In this study, we identified vascular endothelial growth factor (VEGF) as a gene product induced by heregulin-β1. The stimulation by heregulin-β1 of breast cancer epithelial cells induced the expression of the VEGF mRNA and protein and its promoter activity. Heregulin-β1 also stimulated angiogenesis in a VEGF-dependent manner. Herceptin, an anti-HER2 antibody inhibited heregulin-β1-mediated stimulation of both VEGF expression in epithelial cells and angiogenesis in endothelial cells. Because the activation of Pak1 and VEGF expression are positively regulated by heregulin-β1, we hypothesized that Pak1 regulates VEGF expression, and hence explored the role of Pak1 in angiogenesis. We provide new evidence to implicate Pak1 signaling in VEGF expression. Overexpression of a kinase-dead K299R Pak1 leads to suppression of VEGF promoter activity, as well as VEGF mRNA expression and secretion of VEGF protein. Conversely, kinase-active T423E Pak1 promotes the expression and secretion of VEGF. Furthermore, expression of the heregulin-β1 transgene, HRG, in harderian tumors in mice enhances the activation of Pak1 as well as expression of VEGF and angiogenic marker CD34 antigen. These results suggest that heregulin-β1 regulates angiogenesis via up-regulation of VEGF expression and that Pak1 plays an important role in controlling VEGF expression and, consequently, VEGF secretion and function.

Heregulin-␤1 promotes the activation of p21-activated kinase 1 (Pak1) and the motility and invasiveness of breast cancer cells. In this study, we identified vascular endothelial growth factor (VEGF) as a gene product induced by heregulin-␤1. The stimulation by heregulin-␤1 of breast cancer epithelial cells induced the expression of the VEGF mRNA and protein and its promoter activity. Heregulin-␤1 also stimulated angiogenesis in a VEGFdependent manner. Herceptin, an anti-HER2 antibody inhibited heregulin-␤1-mediated stimulation of both VEGF expression in epithelial cells and angiogenesis in endothelial cells. Because the activation of Pak1 and VEGF expression are positively regulated by heregulin-␤1, we hypothesized that Pak1 regulates VEGF expression, and hence explored the role of Pak1 in angiogenesis. We provide new evidence to implicate Pak1 signaling in VEGF expression. Overexpression of a kinase-dead K299R Pak1 leads to suppression of VEGF promoter activity, as well as VEGF mRNA expression and secretion of VEGF protein. Conversely, kinase-active T423E Pak1 promotes the expression and secretion of VEGF. Furthermore, expression of the heregulin-␤1 transgene, HRG, in harderian tumors in mice enhances the activation of Pak1 as well as expression of VEGF and angiogenic marker CD34 antigen. These results suggest that heregulin-␤1 regulates angiogenesis via up-regulation of VEGF expression and that Pak1 plays an important role in controlling VEGF expression and, consequently, VEGF secretion and function.
Growth factors and their receptors play an essential role in the regulation of cancer cell proliferation, metastasis, and angiogenesis. Abnormalities in the expression and actions of human epidermal growth factor receptor (HER) 1 family members contribute to the progression and maintenance of the malignant phenotype and to increased metastasis in human breast cancer (1). HER family receptors are transactivated by recep-tor-receptor interaction in a ligand-dependent manner (2) and thus can utilize more than one pathway to transduce their biological activities. For example, HER-3 and HER-4 receptors bind to more than a dozen isoforms of the heregulins, and each can activate the HER-2 receptor via interactions among heterodimeric interactions (3)(4)(5)(6). Accordingly, an anti-HER2 mAb, Herceptin, has been shown to block heregulin-triggered interactions between HER3 and HER2 interactions, and some of the biological effects of heregulin-␤1 (HRG) (7,8).
The physiologic significance of HRG was established through targeted deletion studies in which mice lacking HRG had developmental abnormalities in the nervous and cardiovascular systems (9,10). The most prominent cardiovascular abnormalities included a lack of the endocardial cushion, which requires mesenchymal cell growth for its development. Heregulin supports angiogenesis (11). Accumulating evidence suggests that HRG, a paracrine growth factor secreted from mesenchymal cells, regulate the progression of breast cancer cells to the invasive phenotype (12)(13)(14). Recently, we confirmed that in the absence of HER-2 overexpression, HRG-stimulated activation of breast cancer cells promotes the development of more aggressive phenotypes; this development includes the formation of lamellipodia and an increase in cell motility through activation of p21-activated kinase-1 (Pak1) (7). Additional evidence that Pak1 may play a role in HRG-mediated invasion of breast cancer cells was demonstrated by the use of kinase-dead Pak1 mutants that promoted cell spreading and the stabilization of focal points, reducing cell invasion (15,16). Despite the widely acknowledged role of HRG in angiogenesis and breast cancer progression, the molecular mechanism by which HRG affects angiogenesis and the potential role that Pak1 signaling plays in angiogenesis remain poorly understood.
Several recent studies suggest that tumor growth and progression are closely linked to angiogenesis. The neovascularization or angiogenesis, which provides nutrient flow to solid tumors, is critical to the initial rapid tumor growth and the growth of metastases (17). The onset of tumor angiogenesis depends on the production of angiogenic factors by the tumor cells or tumor microenvironment that stimulate host organ vascular endothelial cell growth and chemotaxis (18). The signals that influence the expression of angiogenic molecules in tumor cells may come from other tumor cells or from surrounding stromal cells. One of the best characterized angiogenic protein is vascular endothelial growth factor (VEGF), also known as vascular permeability factor (17)(18)(19). VEGF is a potent, selective endothelial cell mitogen and chemotactic protein that is regulated by a number of cytokines, growth factors, and oncogenes (20,21), as well as by phosphatidylinositol 3-kinase signaling (22). Furthermore, overexpression of HER1 and HER2 receptors in human tumor cell lines is associated with increased expression of VEGF and angiogenesis (23).
Overexpression of HER2 occurs in 20 -25% of breast cancer patients. The remainder of breast cancer patients, who have no overexpression of HER-2 receptor levels, often exhibit metastasis. Thus, it is important to explore the potential regulation of VEGF by mesenchymal growth factor HRG in the absence of HER2 overexpression. The results of this study show that HRG regulates the expression of the VEGF promoter, VEGF mRNA, and VEGF protein and its secretion and that HRG stimulates angiogenesis in a VEGF-dependent manner. Our results implicate Pak1 signaling in HRG regulation of VEGF expression. Using dominant-negative or dominant-active Pak1 mutants, we discovered that Pak1 activity was required for the transcriptional expression of VEGF and thereby influenced the level of VEGF secretion.
Metabolic Labeling and Immunoprecipitation of VEGF-Breast cancer cells were grown to 50% confluency and then incubated in serumfree conditions for 48 h. Cells were metabolically labeled with 20 Ci/ml of [ 35 S]methionine for 24 h in a methionine-free medium containing 2% dialyzed fetal bovine serum in the absence or presence of HRG (50 ng/ml) (13). Conditioned media with equal trichloroacetic acid perceptible counts were immunoprecipitated with the desired or control antibody, resolved on SDS-PAGE, and analyzed using autoradiography.
Measurement of VEGF Protein-Breast cancer cells were cultured under serum-free conditions for 24 h and treated in the presence or absence of HRG (50 ng/ml); conditioned medium was harvested after 24 h. The amount of secreted VEGF was measured in the conditioned medium with a VEGF-specific sandwich-ELISA assay (R & D Systems) according to the manufacturer's instructions. VEGF protein levels were normalized to the number of cells.
Northern Analysis-Total cytoplasmic (RNA 20 g) was analyzed by Northern blot analysis. Northern blots were probed with 32 P labeled 473 bp VEGF cDNA fragment (provided by Dr. Abraham Judith). Glyceraldehyde 3-phosphate dehydrogenase levels were used to assess the integrity of the RNA and to control for RNA loading (25).
VEGF Promoter-Reporter Assays-Subconfluent cells cultured in sixwell plates were transiently cotransfected with 2 g of full-length VEGF promotor-luciferase reporter construct (26) and 20 ng of ␤-galactosidase using Fugene-6 reagent according to manufacturer's protocol (Roche Molecular Biochemicals). Twenty-four hours after transfection, cells were serum-starved for 24 h and then treated with or without HRG for 24 h. Cells were lysed with passive lysis buffer, and a luciferase assay was performed using a luciferase reporter assay kit (Promega, Madison, WI). ␤-Galactosidase activity was used to normalize the transfection.
In Vitro Kinase Assay-The Pak1 kinase assay was performed as described (7). Briefly, MCF-7 cells expressing HA-tagged T423E Pak1 were treated with doxycycline (1 g/ml) for 24 h to induce the express kinase-active Pak1, and T423E Pak1 was immunoprecipitated with HA monoclonal antibody. Immunocomplexes were washed three times with lysis buffer and two times with kinase buffer containing 20 mM HEPES (pH 7.4), 1 mM dithiothreitol, 10 mM MnCl 2 , and 10 mM MgCl 2 . The kinase reaction was carried out in kinase buffer supplemented with myelin basic protein as a substrate and 10 ci of [␥ 32 P]ATP at 30°C for 30 min. The kinase reaction was terminated by the addition of 5ϫ SDS-PAGE sample buffer followed by autoradiography. To study Pak kinase activity in tissues from transgenic mice, endogenous Pak1 was immunoprecipitated from tissues lysates using Pak1 antibody (Santa Cruz Biotechnology).
In Vitro Angiogenesis Assay-In vitro angiogenesis in collagen gels was quantitated using spheroids of microvascular endothelial cells (27). To generate these spheroids, 1 ϫ 10 4 cells were suspended in culture medium and seeded in 96-well plates coated with 0.5% (w/v) agarose. HMVEC-L spheroids were generated overnight and then embedded into collagen gels. A collagen stock solution was prepared prior to use by mixing acidic collagen extract of rat tails (equilibrated at 4°C to 3 mg/ml) with 10ϫ Dulbecco's modified Eagle's medium (8:1) and ϳ1 volume of 0.1 N NaOH to adjust the pH to 7.4. A 0.5-ml sample of this stock solution was mixed at room temperature with 0.5 ml of medium containing 10% fetal calf serum and 50 ng/ml HRG in the presence or absence of 100 nM Herceptin or 1 g/ml anti-VEGF antibody. The spheroid-containing gel was rapidly transferred into a 24-well plate and allowed to polymerize, after which 0.15 ml of medium was pipetted on top of the gel. The gel was incubated at 37°C in a 5% CO 2 incubator and 100% humidity.
For the tube formation assay (27), 10 mg/ml of Matrigel was coated in 24-well plates at 4°C. Endothelial cells were gelled at 37°C and then seeded at 5 ϫ 10 4 cells/well onto the Matrigel layers contained 50 ng/ml HRG in the presence or absence of 100 nM Herceptin or 1 g/ml anti-VEGF antibody.
Migration and Chemoinvasion Assays-The rates of migration of HUVEC and HMVEC-L were determined using a modified Boyden chamber assay. Aliquots of HUVEC and HMVEC-L (10 5 cells/aliquot) were placed into the upper chamber of a filter with an 8-m pore size (Costar Transwell) well. The filter inserts with cells were placed in wells of a 24-well culture plate containing 600 l of medium alone, as control, or medium plus 50 ng/ml human recombinant VEGF or 50 ng/ml HRG in the presence or absence of 1 g/ml anti-VEGF antibody, 100 nM Herceptin, or control IgG. After 5 h of incubation at 37°C, nonmigrating cells were scraped off, and the cells that had migrated to the lower surface of the filter inserts were fixed with 100% methanol for 10 min and stained with hematoxylin-eosin. Six randomly selected fields on each filter were then counted.
To test the invasion behavior of HRG-treated MDA-MB435 cells, 8-m filters were coated with Matrigel (20 g/filter). The coated filters were placed in the Boyden chambers. Cells (10 5 ) suspended in Dulbecco's modified Eagle's medium containing 0.1% bovine serum albumin were added to the upper chamber. Conditioned medium from mouse fibroblast NIH3T3 cells was used as a source of chemoattractant and was placed in the lower compartment of the Boyden chambers. After 24 h of incubation, the invaded cells are counted as described above.
Murine Angiogenesis Assay-Angiogenesis was assayed in terms of the growth of blood vessels from subcutaneous tissue into a solid gel of reconstituted basement membrane (i.e. Matrigel) containing the test sample as described (28). Matrigel rapidly solidified at body temperature, thereby trapping the factor, assuring its slow release, and prolonging exposure of surrounding tissues to it. After 2 weeks, mice were killed, and Matrigel plugs were photographed.
Transgenic Studies-A breeding pair of HRG transgenic mice was kindly provided by Dr. Philip Leder (29). Genotype of the animals was confirmed by Southern blotting of tail DNA. Animal breeding and maintenance was performed according to IACUC guidelines. Approximately 50% of the transgenic offspring showed hyperplasia of the harderian gland as reported earlier (29). The hyperplastic harderian gland from transgenic lines and normal harderian gland from wild type animals was dissected and processed for RNA extraction using the TRIZOL method. Expression of HRG was analyzed by RT-PCR. HRG primers were as follows: forward, 5Ј-ATGTCTGAGCGCAAA GAAGGCAGA-3Ј; reverse, 5Ј-TTGCTGATCACTTTGCACATATAC-3Ј. Expression of VEGF was analyzed by RT-PCR followed by Southern blotting. VEGF primers were: forward, 3Ј-CGCGGATCCAGGAGTACCCTGATGAG-5Ј; reverse, 5Ј-CCGGAATTCACATTTGTTGTGCTGT-3Ј.
Sections of the harderian glands were stained with CD34 monoclonal antibody (NeoMarkers, CA). CD34 is a differentiation antigen expressed by most endothelial cells in vivo; expression of CD34 has been shown to be up-regulated by endothelial cells during angiogenesis (30,31). Sections were deparaffinized and rehydrated. Endogenous peroxidase was inactivated with 3% H 2 0 2 , washed in phosphate-buffered saline followed by preincubation with goat serum for 30 min, and then incubated with a mouse monoclonal antibody against CD34 (dilution 1:50) for 60 min. After three washes, the sections were incubated with biotinylated anti-mouse IgG antibodies for 10 min, washed, and incubated with streptavidin-peroxidase for 10 min before the addition of 3-amino-9-ethyl-cabazole. Meyer's hematoxylin was used for counter-staining.

Enhancement by HRG of VEGF Protein Production in Epi-
thelial Cells-During an earlier study to characterize the nature of secreted proteins from HRG-stimulated [ 35 S]methionine-labeled MCF-7 cells, we discovered that HRG enhanced the accumulation of several newly synthesized proteins in the conditioned medium, including a protein with a molecular mass of approximately 20 kDa (Fig. 1A, arrowhead). In the present study, we investigated whether this secreted 20-kDa protein is a VEGF. Immunoprecipitation of the conditioned medium from control (Ϫ) and HRG-stimulated (ϩ) labeled MCF-7 with an anti-VEGF mAb revealed that HRG enhanced the secretion of a 20-kDa VEGF protein (Fig. 1B, arrow). To understand the correlation between VEGF production and the invasiveness of cells, we examined the effect of HRG on the secretion of VEGF and the invasiveness of highly metastatic MDA-MB435 cell (15). Consistent with the invasive nature, MDA-MB435 cells secreted higher levels of VEGF than did MCF-7 cells (Fig. 1C) and were more invasive than MCF-7 cells. As shown in Fig. 1C, however, MDA-MB435 cells responded to HRG with further induction of a 20-kDa VEGF

FIG. 3. Anti-HER receptor antibodies block heregulin-stimulation of VEGF promoter activity.
Breast cancer cells were transiently transfected with a VEGF promoter or control vector and ␤1galactosidase. Some cultures were treated with HRG (50 ng/ml) in the absence (Con, control) or presence of Herceptin (100 nM), C225 (100 nM), anti-HER3 mAb (4 g/ml), or IgG (100 nM) for an additional 24 h. Cells were harvested, and luciferase activity was determined. The results are representative of three to five experiments.
protein band similar to that of VEGF 165 and a weaker induction of a protein similar to VEGF 121 (32). HRG treatment also promoted the ability of MDA-MB435 cells to invade through a porous membrane as measured by a Boyden chamber assay (Fig. 1D). Quantitation of the VEGF in the conditioned medium using a very sensitive ELISA assay confirmed that HRG stimulation of MCF-7 cells tripled the level of VEGF protein over 24 h. The amount of VEGF protein secreted into the conditioned medium from MDA-MB435 cells was about four times higher than from MCF-7 cells, but HRG treatment further increased the level of secreted VEGF to a level 1.7 times the level of VEGF present in the conditioned medium of untreated control cells (Fig. 1E).
HRG-stimulated Increase in VEGF mRNA Expression in Epithelial Cells-We performed a Northern blot analysis to determine whether the HRG-mediated increase in the level of expression of VEGF protein was accompanied by an increase in the level of expression of VEGF mRNA. VEGF mRNA level increased in response to HRG stimulation. HRG treatment continuously increased the steady-state levels of VEGF mRNA by 2-6-fold at 1-9 h after treatment ( Fig. 2A). Treatment of cultures with actinomycin D, an inhibitor of transcription, completely inhibited the HRG-mediated induction of VEGF mRNA (Fig. 2B). To study translational regulation, we utilized cycloheximide, a translational inhibitor. Treatment of cells with cycloheximide stabilized the levels of VEGF mRNA expression; treatment with HRG, however, superinduced the expression of VEGF mRNA (Fig. 2B). Together, these results suggest that HRG regulate VEGF at a pretranslational level.

Effect of Anti-HER-blocking Antibodies on HRG-mediated
Stimulation of VEGF Promoter Activity-To further confirm the role of HRG in the transcriptional regulation of the VEGF gene, cells were transiently transfected with a chimeric luciferase gene fused with the 5Ј region of the VEGF promoter (26), and the activity of the promoter was assayed in the presence or absence of HRG. HRG treatment stimulated VEGF promoter activity in three breast cancer cell lines, all of which had normal levels of HER receptors (Fig. 3). The observed HRGmediated up-regulation of VEGF promoter activity in MCF-7 and MDA-MB435 cells was effectively suppressed by pretreating the cells with the anti-HER-2 mAb Herceptin, which inhibits the formation of HER-3/HER-2 interaction (7), and with anti-HER-3 mAb, which competes with the HRG binding site on the HER-3 receptor (25,33). Anti-epidermal growth factor receptor mAb C255 was more effective than MCF-7 cells in blocking the stimulatory effect of HRG on the VEGF promoter in MDA-MB-231 cells, suggesting that C225 may also prevent HRG-induced interactions between HER-3 and HER-1 receptors. This idea is supported by results showing that anti-HER3 antibody (Ab3) was more inhibitory than Herceptin in MDA-MB231 cells (Fig. 3B) and because MDA-MB231 cells expressed significantly more HER-1 than is present on MCF-7 and MDA-MB435 cells (34).
Effect of p21-activated Kinase-1 Signaling on Expression and Secretion of VEGF-Recent studies from this laboratory showed that Pak1 plays a role in HRG-mediated stimulation of leading edge formation and invasiveness of noninvasive breast cancer cells (7) and in the maintenance of motile/invasive phenotypes of MDA-MB435 cells (15). Because the activation of Pak1 (7) and VEGF (this study) expression are positively regulated by heregulin-␤1 and because VEGF expression can be regulated by phosphatidylinositol 3-kinase (22), a HRG-inducible kinase that is the upstream regulator of Pak1 (7), we hypothesized that Pak1 regulates VEGF expression and therefore explored the role of Pak1 in angiogenesis. To test this possibility, we examined the effect of the dominant-negative K299R Pak1 mutant (7, 15) on HRG-induced activation of the VEGF promoter. Transient co-transfection of MCF-7 cells with pGL3-VEGF and kinase-dead K299R Pak1, but not with control vector, completely suppressed the ability of HRG to stimulate transcription from the VEGF reporter system (Fig. 4A). To further validate the involvement of Pak1 signaling on the level of VEGF expression, we next used well characterized MDA-MB435 cells expressing either K299R Pak1 mutant (435-K16 and 435-K17 cells) or control vector (435-CMV cells) (15). The 435-K17 cells represent a different clone with properties similar to 435-K16 cells. The inhibition of Pak1 signaling was accompanied by a significant reduction (50 -60%) in the level VEGF promoter activity (Fig. 4B) and VEGF mRNA expression (Fig. 4C) and by a 35% reduction in secretion of VEGF protein in the conditioned medium (Fig. 4D) compared with the levels in the control cells. The observed discrepancy between a significant reduction of VEGF mRNA level (60% compared with control cells) and a modest suppression of VEGF accumulation in the conditioned medium (35% compared with control cells) could be due to the possible detection of more than one isoform of VEGF using a radioimmunoassay kit. However, HRG treatment was not able to rescue or induce the observed reduction in the basal VEGF secretion, suggesting the possible requirement of a functional Pak1 pathway during HRG-mediated up-regulation of VEGF.
Effect of Expression of Kinase-active T423E Pak1 on Expression and Secretion of VEGF-To study the effect of dominantactive Pak1 on the expression of VEGF, we used MCF-7 clones expressing kinase-active T423 Pak1 under the control of an inducible tetracycline promoter (16). As expected, expression of HA-tagged T423E-Pak1 increased Pak kinase activity compared with the level in vector control cells (Fig. 5A). Interestingly, the increased expression of kinase-active T423E Pak1 in MCF-7 cells was accompanied by increased VEGF promoter activity (Fig. 5B) and enhanced secretion of VEGF protein in the conditioned medium, as determined by immunoprecipitation of 35 S-labeled VEGF (Fig. 5C) and the results of the ELISA assay (Fig. 5D). Taken together, these results suggest that Pak1 signaling regulates the expression of VEGF and that Pak1 may be an important mediator of HRG regulation of VEGF expression.
Stimulation by HRG of Vascular Tube Formation, Endothelial Cell Migration, and Angiogenesis in Vivo-We tested the ability of HRG to stimulate angiogenesis in cells embedded in collagen gels. In the presence of HRG or VEGF, capillary sprouts originated from HMVEC-L spheroids embedded in collagen gels in the presence of 10% fetal calf serum (Fig. 6A). Inclusion of anti-HER-2 receptor mAb Herceptin or anti-VEGF mAb (but not control antibody; data not shown) suppressed sprouting of vascular endothelial cells in HRG-treated cultures (Fig. 6A).
Because basement membrane can stimulate differentiation, HUVEC were also plated onto a gel composed of reconstituted basement membrane proteins, to which HUVEC attach rapidly. After 24 h, elongated processes appeared in the control cells that had been cultured in the absence of serum. In contrast, the addition of HRG and VEGF to HUVEC promoted the formation of networks of branching and anastomosing cords of cells (commonly known as tube formations) (Fig. 6B). This effect of HRG was significantly blocked by Herceptin or anti-VEGF mAb (Fig. 6B) but not by the control antibody (data not shown).
The effect of HRG on the migration of HUVEC and HM-VEC-L was analyzed using the Boyden chamber assay. Endothelial cells migrated in response to a chemotatic gradient of VEGF and HRG. In this assay, HUVEC were more responsive to HRG than were HMVEC-L. As expected from Fig. 6, A and B, HRG tripled the migration rate of HUVEC and doubled that of HMVEC-L cells (Fig. 6C). However, anti-VEGF mAb and Herceptin each suppressed HRG stimulation of cell migration (Fig. 6C).
The angiogenic activity of HRG was also tested in a murine angiogenesis model involving subcutaneous injection of Matri- gel and HRG. Matrigel rapidly solidified at body temperature, thereby trapping the factor, assuring its slow release, and prolonging exposure of surrounding tissues to it. Two weeks after injection, the Matrigel plugs and the surrounding tissues were histologically examined. In these studies, we used basic fibroblast growth factor as a positive control. On gross examination, the control plugs appeared white; plugs containing HRG appeared bright red, and many contained superficial blood vessels; the effect of HRG on angiogenesis was confirmed by hematoxylin-eosin staining of frozen sections from matrigel plugs (Fig. 7A). These HRG effects were undetectable in plugs containing HRG and anti-VEGF antibody (data not shown). Together, these results suggested that HRG is a very potent angiogenic factor and that VEGF may mediate the angiogenic effects of HRG. In this context, Russell et al. (11) have recently shown that a distinct isoform of heregulin-␤3 can also induce angiogenesis. However, anti-VEGF Ab did not inhibit the effects of heregulin-␤3 on HUVEC (from a donor), suggesting the existence of biological differences between the heregulin-␤1 and -␤3 gene products of the two distinct gene (35).
HRG Regulates Pak1 Activity, VEGF Expression, and Angiogenesis in a Transgenic Model-To evaluate the HRG modulation of Pak1 activity and angiogenesis in vivo, we used an murine mammary tumor virus-driven HRG transgenic model that develops mammary adenocarcinomas and harderian tumors (29). Because harderian tumors can usually be detected by 3 weeks of age as opposed to the 12-16 months required for detection of mammary gland tumors, we used the harderian tumors to establish the proof-of-principle of our hypothesis of HRG regulation of angiogenesis in vivo. Harderian glands from wild type (WT) and HRG-transgenic (HRG-TG) mice were analyzed by RT-PCR for the expression of HRG and by RT-PCR followed by Southern blotting for VEGF. HRG-transgenic mice have significantly elevated levels of HRG transcript compared with wild type mice (Fig. 7B). Interestingly, overexpression of HRG in harderian tumors was accompanied by increased Pak1 kinase activity (Fig. 7D) as well as VEGF expression (Fig. 7C). Furthermore, we also discovered a significant increase in angiogenesis as determined by immunostaining with CD34 antigen (a specific angiogenic marker, Refs. 30 and 31; shown by brown stain) in the harderian gland tumors as compared with wild type tissues (Fig. 7E). Together, these results imply a close relationship among the expression of HRG, Pak1 activity, VEGF expression and angiogenesis in vivo.
In summary, the results presented here demonstrate that: 1) HRG regulates the expression and secretion of VEGF from breast epithelial cancer cells; 2) Pak1 signaling is an important regulator of VEGF expression and functions; 3) HRG stimulates angiogenesis; and 4) VEGF may play a role in the angiogenic effects of HRG on endothelial cells. These data provide evidence that up-regulation of the VEGF pathway in breast cancer epithelial cells by HRG, which is secreted from mesenchymal cells, may have functional implications for the enhancement of invasiveness of breast cancer cells. Any potential upregulation of VEGF by HRG in breast cancer epithelial cells is likely to sustain and perhaps further promote the ability of tumor cells to survive and metastasize by supporting neovascularization. In addition, because HRG is a potent angiogenic factor (this study) and receptors for HRG are also present on endothelial cells (11), we propose a model in which HRG interacts with both mammary epithelial tumor cells and endothelial cells and in which induced secretion of VEGF from tumor cells may further amplify the angiogenic responses of endothelial cells to HRG. FIG. 7. HRG regulation of angiogenesis, and Pak1 activity and VEGF expression in in vivo murine angiogenesis assay. A, matrigel mixed with HRG was injected subcutaneously into mice. Basic fibroblast growth factor was used as a positive control. After 2 weeks, mice were sacrificed, and Matrigel plugs were excised and fixed. Panel I, control (Con); panel II, Matrigel mixed with HRG (5 g); panel IV, fibroblast growth factor (2 g) used as a positive control; panel V, Matrigel mixed with HRG (5 g) and Herceptin (15 g); panels III and VI show the enlargements of the boxed areas in panels II and V, respectively. The data illustrated are representative for one of two independent experiments. B-E, HRG regulation of VEGF expression, Pak1 kinase activity, and angiogenesis in a transgenic model. B, RT-PCR analysis of the expression of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and HRG in wild type (WT) and HRG-transgenic (HRG-TG) harderian tumors from several animals. C, expression of VEGF by RT-PCR followed by Southern blotting. D, Pak1 kinase activity in tissue lysates. E, angiogenesis in HRG-transgenic mice. HRG induces angiogenesis in the harderian gland tumors of murine mammary tumor virus-HRG transgenic mice. Much more enlarged blood vessels were seen in the space between the tumoric glandular tubules of transgenic mice than from that of the normal gland from wild type mice. Panels I and II, immunohistochemical staining of CD34 antigen; the arrows show brown staining of CD34 on endothelial cells. Panels III and IV, hematoxylin-eosin staining.