Promotion of Angiogenesis by Ps20 in the Differential Reactive Stroma Prostate Cancer Xenograft Model1

[CANCER RESEARCH 63, 5859–5865, September 15, 2003] Promotion of Angiogenesis by ps20 in the Differential Reactive Stroma Prostate Cancer Xenograft Model1

Stephanie J. McAlhany, Steven J. Ressler, Melinda Larsen,2 Jennifer A. Tuxhorn,3 Feng Yang, Truong D. Dang, and David R. Rowley4 Department of Molecular and Cellular Biology [S. J. M., S. J. R., J. A. T., F. Y., T. D. D., D. R. R.] and Cell and Molecular Biology Program [M. L.], Baylor College of Medicine, Houston, Texas 77030

ABSTRACT gresses with the developing cancer foci (1). This reactive stroma is composed primarily of the myofibroblast cell type, and exhibits ECM5 Human prostate cancer is associated with a reactive stroma typified by remodeling and gene expression similar to a wound repair type stroma an increase in the proportion of myofibroblast type cells and elevated (1). The DRS xenograft model was developed to examine the effects synthesis of extracellular matrix proteins. Increased vascular density has been identified in the reactive stroma compartment adjacent to both of different prostate stromal cells and stromal gene expression on precancerous and cancerous prostate lesions. The differential reactive progression of LNCaP human prostate cancer cell xenografts in nude stroma (DRS) prostate cancer xenograft model has been developed to mice. These studies have shown that reactive stroma is tumor pro- investigate the role of reactive stroma in prostate cancer progression. moting and that a key early biological action of reactive stroma is the Using this model, we have shown that human prostate stromal cells promotion of angiogenesis (2). Tumors generated in the absence of promote angiogenesis and growth of LNCaP human prostate carcinoma human prostate stroma were significantly delayed in establishing new cell tumors, and that these increases are transforming growth factor blood vessels, resulting in more heterogeneous tumors with ill-defined ␤ (TGF) 1 regulated. Our laboratory isolated and identified previously the vascularity and the appearance of blood lakes in lieu of vessels. In ps20 protein (WFDC1 gene) as a prostate stromal cell secreted protein. The ps20 protein contains a whey acidic protein-type four-disulfide core contrast, DRS tumors constructed with some, but not all, of the human domain, which is a functional motif characterized by serine protease prostate stromal cell lines exhibited angiogenesis as early as day 4 inhibition activity in a number of whey acidic protein domain-containing with extensive vascularity by day 10, illustrating the differential proteins. In the present study, we show ps20 expression by normal human biology of human prostate stroma. In similar DRS tumors, ␤-galac- prostate stromal smooth muscle cells and vascular smooth muscle cells tosidase expressing stromal cells assumed a pericyte position in newly indicating a possible role of ps20 in vessel wall biology. Using in vitro formed capillaries, suggesting a direct interaction with developing assays, we show that ps20 promotes endothelial cell motility but has no vessels (2). In addition, we have reported that the promotion of effect on endothelial cell proliferation. To address the potential effects of angiogenesis in DRS tumors by stroma was TGF-␤1 regulated (3). ps20 in a tumor microenvironment, we used the DRS model to evaluate These studies have directed our interest to factors that specifically both angiogenesis and tumorigenesis of tumors generated under either ps20 or control conditions. DRS tumors generated with LNCaP and regulate new vessel formation and maintenance. human prostate stromal cells in the presence of ps20 showed a 67% Our previous studies have reported that the ps20 protein is ex- increase in microvessel density compared with control tumors. Elevated pressed by prostate stromal cells with strong expression in vessel wall DRS tumor growth in the ps20-treated tumors was reflected by a 29% smooth muscle cells (4). Originally, we characterized ps20 as a increase in wet weight and a 58% increase in volume compared with growth inhibitor of epithelial cells in vitro and purified ps20 to controls. Similar tumors composed of GeneSwitch-3T3 cells engineered to homogeneity from the conditioned medium of a fetal rat urogenital express ps20-V5-His under mifepristone regulation showed a 129% in- sinus mesenchymal cell line (5, 6). Subsequently, we cloned and crease in microvessel density after induction of ps20-V5-His. GeneSwitch- characterized the cDNA encoding rat ps20 (4), human ps20 cDNA, 3T3 cells expressing ps20-V5-His were localized to vessel walls in a mural and genomic DNA encoding the gene we termed WFDC1, as well as cell (pericyte) position indicating a possible direct stabilizing interaction with endothelial cells. In addition, we show that ps20 mRNA synthesis is genomic clones containing mouse ps20 (Wfdc1; Ref. 7). These studies induced by TGF-␤1, a known regulator of endothelial cell-pericyte inter- also localized WFDC1 to human chromosome 16q24 (7). In addition, actions and of stromal cell-induced angiogenesis in DRS tumors. These Rodriguez-Rey et al. recently cloned and submitted a chicken ps20 findings suggest that ps20 may be a TGF-␤1-induced regulator of angio- cDNA to the GenBank/European Molecular Biology Laboratory da- genesis that functions by either promoting endothelial cell migration or by tabase (AJ438920). There is high conservation of ps20 between spe- contributing to pericyte stabilization of newly formed vascular structures. cies, ranging from 61% to 96% identity of the mature protein. Each ps20 protein sequence contains a functional motif identified INTRODUCTION as a WAP type-four disulfide core domain consisting of eight cys- teines forming four characteristically paired disulfide bonds. The Prostate cancer is typified by the coevolution of a reactive stroma, domain and protein family is named after the first characterized which initiates during prostatic intraepithelial neoplasia and pro- member, WAP (8). The WAP domain confers a serine protease inhibition activity to several family members including elafin and Received 4/24/03; revised 6/25/03; accepted 6/30/03. SLPI (9, 10). Elafin is a potent inhibitor of elastase and proteinase 3, The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with each of which are elastin-degrading proteases (11). SLPI inhibits a 18 U.S.C. Section 1734 solely to indicate this fact. broader range of serine proteases including trypsin, chymotrypsin, 1 Supported by NIH Grants RO1-CA58093, RO1-DK45909, Specialized Programs of elastase, and cathepsin G (11). Both elafin and SLPI are secreted Research Excellence CA58204, and UO1-CA84296. S. J. M. is a Hudson Scholar of the Baylor College of Medicine Medical Scientist Training Program. proteins known to protect host tissues from the effects of excessive 2 Present address: Craniofacial Developmental Biology and Regeneration Branch, proteolysis and to promote appropriate tissue repair in numerous National Institute of Dental and Craniofacial Research, NIH, 30 Convent Drive, MSC 4370, Bethesda, MD 20892-4370. 3 Present address: Wyle Laboratories, Life Sciences Systems and Services, 1290 5 The abbreviations used are: ECM, extracellular matrix; DRS, differential reactive Hercules Drive, Houston, TX 77058. stroma; BAE, bovine aortic endothelial; HUVEC, human umbilical vein endothelial cell; 4 To whom requests for reprints should be addressed, at Department of Molecular and ps20, prostate stromal 20 kDa; SLPI, secretory leukocyte protease inhibitor; sm ␣-actin, Cellular Biology, Baylor College of Medicine, Houston, TX 77030. Phone: (713) 798- smooth muscle ␣-actin; TGF, transforming growth factor; WFDC1, whey acidic protein- 6220; Fax: (713) 790-1275; E-mail: [email protected]. type four-disulfide core-1; WAP, whey acidic protein; CHO, Chinese hamster ovary. 5859

model systems (12–14). These data suggest a role for these protease calf serum, 100 units/ml penicillin, 100 ␮g/ml streptomycin, and 0.29 mg/ml inhibitors in reactive stroma biology at numerous tissue sites. On the L-glutamine (Invitrogen). Cultures between passages 10 and 20 were used for basis of the immunolocalization of ps20 to smooth muscle cells and these studies. the presence of a signal peptide with secretion of the native protein HUVECs (Clonetics, Walkersville, IL) were cultured in endothelial cell from cells in vitro, it is theorized that ps20 is secreted by smooth growth medium-2 (Clonetics). Cells between passages 3 and 8 were used in muscle cells in vivo and is a component of the ECM environment, these studies. where it may function as a serine protease inhibitor. LNCaP human prostate carcinoma cells (American Type Culture Collec- The purpose of the present study was to additionally evaluate ps20 tion) were cultured in RPMI 1640 (Invitrogen) supplemented with 10% fetal ␮ expression in human prostate stroma, including vessels, and to deter- bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. mine the effects of ps20 on endothelial cell migration, angiogenesis, HPS-19B human prostate stromal cells were established using our protocols reported previously (6, 17). Briefly, prostate tissue from a 19-year-old organ and tumorigenesis using in vitro assays, and the DRS xenograft donor was obtained from the Baylor College of Medicine Prostate Specialized prostate cancer model. We report that although ps20 is expressed in Program of Research Excellence Pathology and Tissue Microarray Core. A normal prostate stroma, expression was restricted to vessel wall tissue core was isolated from peripheral zone prostate graded as normal by smooth muscle cells in samples of human prostate cancer Gleason 3 histopathological criteria. The core was diced into 1-mm cubes, washed with foci. Additionally, endothelial cell motility was stimulated in the HBSS, and cultured in a 24-well plate with Bfs medium: DMEM containing presence of ps20, whereas endothelial cell proliferation was not sig- 5% fetal bovine serum, 5% Nu-Serum (BD Biosciences, Bedford, MA), 0.5 nificantly altered. The presence of ps20 in the DRS tumor microen- ␮g/ml testosterone, 5 ␮g/ml insulin, 100 units/ml penicillin, and 100 ␮g/ml vironment significantly increased both angiogenesis and tumorigene- streptomycin (Sigma). The explants were cultured at 37°Cin5%CO2, and sis, as determined by microvessel density, tumor wet weights, and culture medium was changed every 48 h. Stromal cells migrated out of the volumes. In addition, we determined that ps20 mRNA expression was tissue, attached to the culture dish, and were grown to confluence. The increased by TGF-␤1, a known regulator of angiogenesis in prostate remaining tissue was removed and the cells passaged by routine procedures. cancer. Together, these data suggest that ps20 is expressed in the Standard immunocytochemistry procedures were used to characterize the vessel wall in prostate cancer and can function to promote angiogen- HPS-19B cells as cytokeratin negative (1–124-161; Boehringer Mannheim, esis, possibly by stimulating endothelial cell migration or by stabili- Indianapolis, IN), and vimentin positive (sc-7557; Santa Cruz Biotechnology, zation of the vessel wall during vessel formation. These studies also Santa Cruz, CA). Approximately 60% of the cells were sm ␣-actin positive, suggest that ps20 may be a target for novel prostate cancer therapeu- similar to our stromal cell lines reported previously (2). Cultures between tics directed toward angiogenesis in the reactive stroma compartment. passage 10 and 15 were used for experiments. Our PS-1 stromal cell line was derived from an adult rat (Harlan Sprague Dawley) prostate organ explant as described previously (17) and maintained in MATERIALS AND METHODS Bfs medium. Preparation of Conditioned Medium. GeneSwitch-CHO-ps20pGENE/ Tissues. Radical prostatectomy specimens were obtained from the Baylor V5-His or GeneSwitch-CHO-pGENE/V5-His control cells were plated and College of Medicine prostate Specialized Program of Research Excellence grown to confluence. Cells were pulsed with 100 pM mifepristone for 24 h, pathology and tissue microarray core. Specimens were processed as described then incubated in Ham’s F-12 medium supplemented with 1% fetal bovine previously (15). Briefly, tissues were cut into 5-mm slices, fixed in 10% serum and 2 mML-glutamine for 48 h. Conditioned medium was collected and neutral buffered formalin, and embedded in paraffin as whole mounts. Whole- cleared by low-speed centrifugation. Approximate concentration of ps20-V5- mount thin sections of 5 ␮m were stained with H&E and evaluated for histological differentiation. For immunostaining, the whole mounts were cut in His in conditioned medium was determined by Western analysis using an half, and thin sections were mounted on standard microscope slides. anti-V5 antibody comparing conditioned medium with serial dilutions of ϳ Cell Lines. COS-1 cells (monkey kidney fibroblast-like; American Type purified insect-expressed ps20-V5-His. Conditioned medium contained 5 Culture Collection, Manassas, VA) were cultured in DMEM (Invitrogen, ␮g/ml ps20-V5-His protein. Carlsbad, CA), containing 10% fetal bovine serum (Hyclone, Logan, UT), 100 Migration Assays. BAE cell migration assays were carried out in transwell units/ml penicillin, and 100 ␮g/ml streptomycin (Sigma, St. Louis, MO). chambers (Costar-Corning Inc., Corning, NY). Polycarbonate filter inserts for COS-1 stable cell lines expressing rat ps20 were generated and maintained as 24-well plates with 8-␮m pores were coated with 17 ␮g/cm2 rat tail collagen reported previously (4). I (BD Biosciences). Confluent BAE cells were incubated in DMEM-Low The GeneSwitch System (Invitrogen) was used to create stable cell lines Glucose with 0.1% BSA (Sigma) for 2 h, then 8.25 ϫ 104 cells in 150 ␮l with mifepristone-inducible expression of a V5-His tagged ps20 protein. medium were plated onto each insert. To each insert, 50 ␮l ps20 conditioned GeneSwitch-CHO and GeneSwitch-3T3 cells expressing the GeneSwitch reg- medium (final concentration of 50 nM ps20-V5-His protein) or control condi- ulatory protein from the pSwitch vector were purchased from Invitrogen. tioned medium were added, whereas each lower chamber received 600 ␮l GeneSwitch-CHO cells were cultured in Ham’s F-12 medium (Invitrogen) DMEM-Low Glucose with 0.1% BSA. Assays were carried out at 37°Cin5% supplemented with 10% fetal bovine serum, 2 mML-glutamine, and 100 ␮g/ml CO2 for 16 h. Cells that had not migrated through the filter were removed, and hygromycin-B (Invitrogen). GeneSwitch-3T3 cells were cultured in DMEM those that had were stained and fixed with crystal violet for 10 min [0.5% (w/v) ␮ containing 10% fetal bovine serum and 50 g/ml hygromycin-B. Rat ps20 crystal violet, 3.2% formaldehyde (Sigma), 0.17% NaCl (Fisher Scientific, cDNA was inserted into the pGENE/V5-His vector to generate ps20pGENE/ Pittsburgh, PA), and 22% ethanol (AAPER Alcohol and Chemical, Co., V5-His, which was transfected in parallel with empty vector controls into both Shelbyville, KY)], washed three times with water, and allowed to dry. From GeneSwitch-CHO and GeneSwitch-3T3 cell lines using FuGENE 6 transfec- each filter, four random fields at ϫ200 were photographed, and an observer tion reagent (Roche, Indianapolis, IN). Stably transfected GeneSwitch-CHO blinded to experimental conditions counted and averaged migrated cells. cells were selected with 400 ␮g/ml Zeocin (Invitrogen), whereas GeneSwitch- 3T3 cells were selected with 300 ␮g/ml Zeocin. Resistant clones from each cell Assays were performed at least three times in triplicate. COS-1 cell migration assays were carried out using standard wounding line were treated with 100 pM mifepristone (Sigma) and screened for fusion ⌬ protein expression by Western analysis of the conditioned medium with assays. The parent cell line COS-1, control line COS-1 pBKCMV- lac-4, and ⌬ anti-V5 antibody (Invitrogen). All of the stably transfected cell lines were experimental lines COS-1 pBKCMV- lac-rps20–3/-4 were plated on glass 4 2 maintained in 300 ␮g/ml Zeocin. Cells lines used in this study represent pooled coverslips at 3.58 ϫ 10 cells/cm and allowed to attach overnight. Triplicate clones of stably transfected cells. scrapes were made in the cell monolayers, and coverslips were harvested at BAE cells (kindly provided by Dr. Karen Hirschi, Departments of Pediatrics multiple time points from 0 to 96 h. Coverslips were stained and fixed with and Molecular and Cellular Biology, Baylor College of Medicine, Houston, crystal violet for 10 min, washed three times with water, and allowed to dry. TX; Ref. 16) were cultured in DMEM-Low Glucose supplemented with 10% Images were analyzed using the public domain NIH Image program (devel- 5860

Downloaded from cancerres.aacrjournals.org on October 5, 2021. © 2003 American Association for Cancer Research. ps20 PROMOTES ANGIOGENESIS oped at the U.S. NIH and available on the Internet)6 with the density of wound adhesion molecule-1) antibody (MEC13.3, rat monoclonal), and biotin-conju- area filled measured in arbitrary units. Experiments were repeated in triplicate gated antirat IgG secondary antibody were purchased from BD PharMingen three times. (San Diego, CA). Anti-V5 antibody was purchased from Invitrogen (46–0705, Proliferation Assay. BAE cells or HUVECs were plated at 3000 cells/cm2 mouse monoclonal). Antigen retrieval was necessary for both antibodies; in 24-well plates and allowed to attach overnight. Cells were incubated in sections were incubated in 0.1% trypsin (Zymed, South San Francisco, CA) for serum-free medium for 24 h. Serum-free medium was removed and replaced 10 min at 37°C (CD31) or subjected to high temperature steamer treatment in with 250 ␮l of DMEM-Low Glucose supplemented with 2% calf serum (BAE 10 mM sodium citrate buffer (pH 6.0) for 20 min (V5). For V5 staining, cells) or endothelial cell growth medium-2 (HUVECs) and 250 ␮l of either sections were incubated in 20 ␮g/ml goat-antimouse Fab fragment (Jackson ps20 (final concentration of 100 nM ps20-V5-His protein) or control condi- ImmunoResearch, West Grove, PA) for 30 min at 37°C to block host immuno- tioned medium. BAE cells were incubated for 48, 72, or 96 h, harvested, and globulins. Sections were incubated with primary antibody overnight at 4°C counted using a hemacytometer. HUVECs were counted at the 48-h time point (CD31, 1:50; V5, 1:200) and then with secondary antibody (antirat, 1:100; only. Each experiment was carried out three independent times in triplicate. Research Genetics Universal Secondary) for 45 min at 37°C (CD31) or 4 min Animals. Athymic NCr-nu/nu male homozygous nude mice, 6–8 weeks of at 50°C (V5). age, were purchased from Charles River Laboratories (Wilmington, MA). Microvessel Density Analysis. Tumor sections were stained for CD31, and Experiments were conducted in accordance with the NIH Guide for the Care four random high-power fields (ϫ400) were photographed for each tumor. and Use of Laboratory Animals. Vessels were counted by a blinded observer using standard criteria (21). For Preparation of Xenograft Tumors. DRS model xenograft tumors were DRS-LNCaP experiments, 12 tumors (48 fields) from duplicate experiments generated under three-way conditions (LNCaP ϩ stromal cells ϩ Matrigel) as were analyzed for each treatment group. For 3T3 cell tumors, 4 tumors (16 described previously (2) with some modifications. Briefly, frozen aliquots of fields) were analyzed for each treatment group at each time point. cells (32 ϫ 106 LNCaP and 8 ϫ 106 HPS-19B cells) were thawed, mixed, and Northern Analysis. ps20 mRNA expression was analyzed by Northern washed with 10-ml culture medium. Cells were pelleted at 400 ϫ g for 2 min, blotting as described previously (4). Briefly, PS-1 cells were incubated in then resuspended in 12 ml culture medium and divided into two 6-ml aliquots. DMEM with 5 ␮g/ml insulin, 5 ␮g/ml transferrin, and 5 ng/ml selenium (ITS; Cells were pelleted and resuspended in 300 ␮l of either ps20 (final concen- Sigma), supplemented with either vehicle control, 25 pM,or50pM TGF-␤1 tration of 80 nM ps20-V5-His protein) or control conditioned medium. Matrigel (porcine; R&D Systems, Minneapolis, MN) for 24 h. Total RNA was prepared ECM (Becton Dickinson, Bedford, MA) was stored in 500-␮l aliquots at using RNA Stat-60 total RNA isolation reagent (Tel-Test, Inc., Friendswood, Ϫ20°C and thawed on ice at 4°C for 3 h before use. Cells were incubated on TX), and electrophoresed through 1% agarose gels containing 1ϫ 4-morpho- ice for 1.5 min, then mixed with Matrigel and drawn into a 1-ml syringe with linepropanesulfonic acid and 5.1% of a 37% formaldehyde solution and a 20-gauge needle. Using a 25-gauge needle, 100 ␮l of cell-Matrigel suspen- transferred to neutral charged nylon membranes (Schleicher & Schuell, Keene, sion (2 ϫ 106 LNCaP and 0.5 ϫ 106 HPS-19B cells; 4:1 ratio) were injected NH). Inserts of ps20 cDNA segments from T340pCRII and 42T7pCRII were 32 s.c. in each lateral flank. Six sites (three animals) were injected per treatment excised, gel-purified, labeled by random priming with [␣- P]dCTP, purified, 6 group for each experiment, and tumors were harvested on day 10. Experiments and denatured. Blots were UV cross-linked and exposed to probe at 1–5 ϫ 10 were carried out four times. counts/ml overnight at 42°C, and washed at 60°C in 0.1% SDS containing ϫ GeneSwitch-3T3-ps20pGENE/V5-His cell tumors were generated in a sim- 2–0.1 SSC. Blots were exposed to film for 24 h to 2 weeks, as necessary. ilar manner. Frozen cells (4 ϫ 106) were thawed and washed twice with 10 ml RNA transferred to the blot was stained with methylene blue as a loading DMEM supplemented with 10% fetal bovine serum. Cells were pelleted and control. Similar results were obtained in duplicate experiments. NIH Image resuspended in a total of 300-␮l wash medium. Matrigel preparation and cell was used to quantitate data, and ps20 mRNA expression was normalized to injections were carried out as described above, resulting in injections of 18S rRNA. 0.5 ϫ 106 stromal cells per site. To induce expression of ps20-V5-His, animals Statistical Analysis. Two-way ANOVA, followed by Bonferroni post received mifepristone (RU486; Sigma) at 0.5 mg/kg administered as 100 ␮l i.p. tests, was used to compare treatment groups over time in both the COS-1 injections at the time of tumor injection and repeated every 48 h until tumors migration and the cell proliferation studies. BAE cell migration, tumor weights were harvested on day 11. Animal dosage of mifepristone was based on and volumes, and tumor microvessel density were each evaluated with an protocols shown previously to induce gene expression in vivo and not affect unpaired t test. Statistical analysis was performed with Prism version 3 Ͻ tumor weight or volume (2, 18, 19). Control mice were treated with vehicle (GraphPad Software, San Diego, CA). Error bars represent SE, and P 0.05 alone. A duplicate experiment with similar results was carried out with tumors was considered statistically significant. harvested on day 15. All of the tumors were weighed and measured in three dimensions using calipers. Tumor volume was calculated with the formula: V ϭ 0.52 ϫ RESULTS length ϫ width ϫ height (20). Tissues were fixed in 4% paraformaldehyde overnight at 4°C and then washed three times with PBS. Tumors were paraffin Evaluation of ps20 Expression in Both Normal Human Prostate embedded, and 5-␮m sections were cut and mounted onto ProbeOn Plus slides Tissue and Prostate Cancer. Previous studies by our laboratory (Fisher Scientific). All of the sections were stained with H&E for histo- have localized ps20 protein expression to the periacinar ring of logical analysis. smooth muscle immediately adjacent to epithelial acini in the rat Immunohistochemistry. Immunostaining for ps20 and sm ␣-actin was prostate gland, as well as to rat vascular smooth muscle (4). In the performed using the EnVision ϩ System for rabbit primary antibodies or current study, positive staining was detected in the prostate stromal mouse primary antibodies (Dako, Carpinteria, CA) following the manufactur- compartment of normal human prostate tissue (Fig. 1A) that consists er’s recommendations. For ps20 staining, our previously characterized affinity predominately of smooth muscle cells (1), as shown by sm ␣-actin purified rabbit polyclonal ps20ab-1 was used (4). Tissue sections were incu- staining (Fig. 1B). Intense ps20 staining was also detected in the bated with ps20ab-1 diluted 1:150 for1hat37°C, and with labeled polymer smooth muscle cell layer of vessel walls (Fig. 1A, arrowheads). To for 45 min at room temperature. For sm ␣-actin staining, we used the mouse date, our staining of human prostate cancer samples has revealed monoclonal 1A4 antibody (Sigma). Tissue sections were incubated with primary antibody diluted 1:5000 for 30 min and labeled polymer for 30 min, both variable degrees of ps20 expression. However, strong ps20 expression at room temperature. was consistently observed in the vascular smooth muscle cells in Immunostaining for CD31 and V5 was performed with the MicroProbe Gleason 3 cancer foci evaluated in this study (Fig. 1C, arrowhead), Staining System (Fisher Biotech) following our protocol published previously whereas the reactive stroma compartment had little to no ps20 immu- (2). Reagents used with this capillary gap system were purchased from Re- noreactivity. These data suggested that ps20 may function in vessel search Genetics (Huntsville, AL). Antimouse CD31 (platelet endothelial cell wall biology. Accordingly, to additionally characterize ps20 function, we examined the effects of ps20 on endothelial cell biology in vitro 6 Internet address: http://rsb.info.nih.gov/nih-image/. and angiogenesis in vivo in a prostate cancer xenograft model. 5861

migrated through the chamber membrane compared with 53 Ϯ 7 control conditioned medium-treated cells (Fig. 2A; P Ͻ 0.05, unpaired t test), representing a 104% increase. To ensure that the active component of the conditioned medium preparation was ps20 protein, we performed wound migration assays with ps20 expressing COS-1 cell lines described previously (4). Two ps20 expressing clonal cell lines, COS-1 pBKCMV-⌬lac-rps20-3 and 4, exhibited increased cell migration and wound closure over time when compared with either the COS-1 parent cell line or control COS-1 cells stably transfected with control vector (Fig. 2B; P Ͻ 0.05, two-way ANOVA with Bonferroni post-test). At 48 h, the average migration of ps20 expressing cell lines showed a 75% increase compared with the average migration of the control cell lines. Our previous reports describe ps20 as a growth inhibitor of the prostate epithelial carcinoma cell line PC-3 (4, 5). Factors involved in angiogenesis often exhibit endothelial cell growth regulating properties. For this reason, we examined the effect of ps20 on BAE cell and HUVEC proliferation in vitro. BAE cells and HUVECs exposed to ps20 showed no statistically significant differences in proliferation when compared with control cells (data not shown). This finding indicates that the ps20-induced increase in the number of migrated cells does not reflect a combined migration and growth stimulation effect. ps20 Stimulates Angiogenesis and Tumor Growth in Vivo. We have shown previously that stromal cells promote tumorigenesis and early angiogenesis in our DRS xenograft tumors composed of Matri- gel, LNCaP human prostate carcinoma cells, and various human

Fig. 1. ps20 protein is expressed by both vascular and prostate stromal smooth muscle cells. A, normal human prostate tissue was stained using ps20ab-1. ps20 protein expression was detected in both vascular smooth muscle (arrowheads) and prostate stromal smooth muscle. B, A serial section of tissue was stained using an antibody to sm ␣-actin, and expression was detected in cell populations equivalent to those positive for ps20 expression; vascular smooth muscle (arrowheads) and stromal smooth muscle cells. C,a focus of Gleason 3 prostate cancer stained for p20 indicates decreased ps20 protein expression in the general stromal compartment compared with stroma surrounding normal epithelial glands, but strong ps20 expression in vascular smooth muscle (arrowhead). Bars,50␮m.

ps20 Stimulates Migration of Both BAE and COS-1 Cells. To Fig. 2. ps20 increases migration of both BAE and COS-1 cells. A, numbers of migrated BAE cells when incubated with either vehicle control or ps20 (50 nM) as assessed using ,ء .(determine whether ps20 affects endothelial cell motility in vitro,we transwell chambers (control n ϭ 3, ps20 n ϭ 3, experiment was repeated three times evaluated ps20-treated and control BAE cells in transwell chamber statistically significant increase in migration of ps20 exposed BAE cells when compared Ͻ assays. WAP family members are difficult to express as biologically with control cells (P 0.05). B, COS-1 cells stably transfected with either ps20 (COS-1 pBKCMV-⌬lac-rps20-3 and COS-1 pBKCMV-⌬lac-rps20-4 ࡗ) or empty vector active proteins in nonmammalian systems (22, 23). Therefore, we (COS-1 pBKCMV-⌬lac ‚), and nontransfected COS-1 cells (Ⅺ) evaluated in an in vitro used mammalian expressed ps20-V5-His from CHO cells secreted wound closure assay to assess cell migration (n ϭ 3 for each cell line at each time point; statistically significant increase in migration of ,ء .(experiment was repeated three times under inducible conditions using the GeneSwitch system. In the COS-1 cells stably expressing ps20 when compared with either control cell line presence of ps20, an average of 108 Ϯ 14 BAE cells/high power field (P Ͻ 0.05). 5862

Fig. 3. Incorporation of ps20 into the DRS xenograft tumor microenvironment increases microvessel density, tumor weight, and volume. A and B, DRS tumors composed of LNCaP cells, HPS-19B stromal cells, Matrigel, and exogenously added ps20 (B,80nM) or vehicle control (A), harvested on day 10, and stained for the endothelial cell marker CD31. Bar, statistically ,ء .(␮m. C, microvessel density, as assessed by CD31 positive structures counted by a blinded observer (n ϭ 40–48 fields, 10–12 tumors, for each treatment group 50 significant increase in microvessel density of ps20 DRS tumors when compared with control tumors (P Ͻ 0.001). D, tumor wet weights for DRS tumors generated in either the presence statistically significant increase in wet weights of ps20 tumors when compared with control tumors ,ء .(of absence of exogenously added ps20 (n ϭ 24 tumors for each treatment group statistically significant increase in volumes of ps20 tumors when compared ,ء .(P Ͻ 0.0001). E, tumor volumes for ps20 and control tumors (n ϭ 24 tumors for each treatment group) with control tumors (P Ͻ 0.0001). prostate stromal cell lines (2, 3). To determine the effect of ps20 on stabilization. Tumors expressing ps20-V5-His (mifepristone-treated) DRS tumor growth and angiogenesis, we mixed either ps20 or control exhibited a 129% elevation in microvessel density relative to control conditioned medium with Matrigel used for injections of LNCaP and tumors, 48 Ϯ 5 and 21 Ϯ 3, respectively (Fig. 4, C–E; P Ͻ 0.0001, HPS-19B human prostate stroma cells. Our previous studies have unpaired t test). shown that day 10 of tumor growth was an optimal time point to Expression of ps20 Is TGF-␤1 Regulated. Previous studies in assess early angiogenesis in DRS tumors (2). Accordingly, cells were our laboratory have shown that stromal induced angiogenesis in the injected and tumors were harvested on day 10. To evaluate tumor DRS model TGF-␤1 regulated (3). Therefore, we assessed TGF-␤1 vasculature, sections were stained with an anti-CD31 (PECAM-1) regulation of Wfdc1 (ps20) gene expression. Northern analysis indi- antibody (Fig. 3A, control, and B, ps20), a marker for endothelial cated a dose-dependent increase in ps20 mRNA expression when cells, and microvessel density was determined following procedures PS-1 rat prostate stromal cells were treated with 25 and 50 pM TGF-␤1 we have reported previously (2, 3). Tumors containing ps20 exhibited (Fig. 5). The ps20 mRNA expression was elevated 4.5- and 11.3-fold a microvessel density of 20 Ϯ 2 compared with 12 Ϯ 2 vessels/high above basal levels with 25 and 50 pM TGF-␤1, respectively. Similar power field in control tumors (Fig. 3C; P Ͻ 0.001, unpaired t test), results with reverse transcription-PCR analysis of ps20 mRNA ex- representing a 67% increase in vessel density. In addition, tumors pression after TGF-␤1 induction were noted (data not shown). containing ps20 exhibited a 29% increase in growth as determined by Ϯ Ϯ wet weights, 22 0.76 mg for ps20 tumors compared with 17 0.99 DISCUSSION mg for control tumors, as well as a 58% elevation in tumor volumes, 19 Ϯ 1mm3 for ps20 tumors compared with 12 Ϯ 1mm3 for control A primary interest of our laboratory is the role of the stromal tumors (Fig. 3, D and E; P Ͻ 0.0001, unpaired t tests). microenvironment in prostate cancer development and progression. In addition to DRS tumors, we investigated whether 3T3 cell The ps20 protein and corresponding gene, WFDC1, were originally tumors engineered to overexpress ps20 would yield similar results. isolated from prostate stromal cells and characterized by our labora- Mice injected with GeneSwitch-3T3-ps20pGENE/V5-His cells were tory (4–7). The ps20 protein contains a WAP-type four-disulfide core treated with either 0.5 mg/kg mifepristone or with vehicle control. domain identifying ps20 as a member of this serine protease inhibitor Tumors harvested from mifepristone-treated animals showed strong family. Whereas our studies characterizing changes in ps20 expres- ps20-V5-His staining, whereas immunoreactivity was not detected in sion during prostate cancer progression are ongoing, we have consis- control tumors (Fig. 4, A control, and B, mifepristone). In addition, we tently noted strong ps20 expression from vascular smooth muscle noted that ps20-V5-His expressing cells in the tumors from the cells. Because of this expression and potential activity as a serine mifepristone-treated mice were often localized immediately adjacent protease inhibitor, we sought to determine the effects of ps20 on to newly formed vessels in a pericyte position (Fig. 4B, inset, arrows). vessel biology, specifically angiogenesis in a tumor microenvi- This suggests a possible direct role in new vessel formation or ronment. 5863

Fig. 4. Analysis of GeneSwitch-3T3- ps20pGENE/V5-His cell tumors. A and B, tumors generated with GeneSwitch-3T3-ps20pGENE/V5- His cells and mice treated with either vehicle control (A) or 0.5 mg/kg mifepristone every 48 h (B) harvested on day 15 and stained for ps20-V5-His expression. Bar, 50 ␮m. B, inset, high magnifica- tion image of ps20-V5-His stained cells (arrows) adjacent to RBC containing vessel. Bar, 10 ␮m. C and D, control (C) and mifepristone induced (D) GeneSwitch-3T3-ps20pGENE/V5-His cell tumors harvested on day 11 and stained for the endothelial cell marker CD31. Bar,50␮m. E, tumor sections were counted for CD31 positive structures by a blinded observer (n ϭ 16 fields, 4 tumors, for each -statistically significant in ,ء .(treatment group crease in microvessel density in ps20-V5-His expressing tumor as compared with control tumors (P Ͻ 0.0001).

As a secreted protein, ps20 may affect several cell types including growth factor from the ECM (25). The growth inhibition activity of endothelial cells. Stimulation of endothelial cell migration in vitro ps20 may rely on a similar indirect mechanism and be significant only often correlates to elevated angiogenesis in vivo (24). We determined in specific cell types under certain physiological conditions. that ps20 significantly increased BAE cell motility compared with A number of human epithelial cancers, including prostate cancer, control cells. In addition, we report that ps20 expression in stably are known to involve a stromal reaction where increases are observed transfected COS-1 cells promotes migration when compared with in stromal myofibroblast type cells, ECM deposition, and angiogen- either the parent COS-1 cell line or to control cells stably transfected esis (26, 27). The current study examined the effect of ps20 on DRS with empty vector. Whereas a specific mechanism of action remains tumors generated with LNCaP human prostate carcinoma cells and to be determined, we hypothesize that ps20 functions as a serine HPS-19B human prostate stromal cells. We determined that ps20 protease inhibitor to stabilize the ECM and alter cell-matrix interac- increased tumor wet-weight, volume, and microvessel density in day tions with a resulting increase in migration. Whether this specific 10 tumors. We also assessed the effect of ps20 on angiogenesis in 3T3 effect accounts for the increase in microvessel density observed here cell tumors to determine whether the ps20-induced increase of angio- is not known. genesis in the DRS model is specific to prostate cancer xenograft Many factors that alter angiogenesis in vivo affect endothelial cell tumors or is a more generalized effect. 3T3 cell tumors with induced proliferation in vitro. The original description of ps20 characterized it expression of ps20 were found to have an increased microvessel as a growth inhibitor of prostate epithelial cells, as well as stably density, similar to the DRS tumors. Together, these results suggest transfected COS-1 cells (4, 5). In the current study, ps20 did not affect that ps20-increased angiogenesis is because of a generalized effect. the growth of BAE cells or HUVECs in culture. The WAP family Because ps20 is a serine protease inhibitor family member, its ob- member elafin is capable of inhibiting cell proliferation by preventing served properties may reflect inhibition of protease activity, and elastase-dependent release of the growth factor basic fibroblast resultant regulation of ECM remodeling and cell-matrix interactions culminating in enhanced endothelial cell migration into developing tumors or stabilization of newly formed vessels. Whereas a suspected protease inhibitor may initially be predicted to inhibit angiogenesis, as many proteases are proangiogenic, there is emerging evidence that some protease inhibitors may stimulate angiogenesis. Plasminogen activator inhibitor 1, a urokinase plasminogen activator serine protease inhibitor, promotes angiogenesis by its protease inhibition properties, and high levels of plasminogen activator inhibitor 1 are predictive of poor survival rates in cancer patients (28–31). Another protease-dependent cascade that affects angiogenesis is the release of antiangiogenic protein fragments from the base- ment membrane. The antiangiogenic factor, endostatin, is generated from collagen XVIII in a cleavage reaction involving the serine Fig. 5. Treatment of stromal cells with TGF-␤1 results in increased ps20 mRNA protease elastase (32). Because of the shared WAP-type serine pro- expression. PS-1 rat prostate stromal cells were treated with vehicle control, 25 pM,or50 tease inhibitor domain, ps20 is related to the potent elastase inhibitor pM TGF-␤1 for 24 h. Total RNA was prepared and ps20 mRNA content was analyzed by Northern blotting. RNA transferred to the blot was stained with methylene blue as a elafin and, thus, may inhibit the production of endostatin-like antian- loading control. giogenic peptides. Accordingly, protease inhibitors may balance or 5864

Downloaded from cancerres.aacrjournals.org on October 5, 2021. © 2003 American Association for Cancer Research. ps20 PROMOTES ANGIOGENESIS control excessive ECM degradation that would otherwise hinder new 11. Schalkwijk, J., Wiedow, O., and Hirose, S. The trappin gene family: proteins defined vessel formation. by an N-terminal transglutaminase substrate domain and a C-terminal four-disulphide core. Biochem. J., 340: 569–577, 1999. In addition to possibly affecting the endothelial cell sprouting and 12. Barolet, A. W., Nili, N., Cheema, A., Robinson, R., Natarajan, M. K., O’Blenes, S., migration phase of angiogenesis, ps20 may function to stabilize newly Li, J., Eskandarian, M. R., Sparkes, J., Rabinovitch, M., and Strauss, B. H. Arterial elastase activity after balloon angioplasty and effects of elafin, an elastase inhibitor. formed vascular structures. Our localization of ps20 secreting stromal Arterioscler. Thromb Vasc. Biol., 21: 1269–1274, 2001. cells to tumor neovasculature indicates that these stromal cells are 13. Zaidi, S. H., You, X. M., Ciura, S., Husain, M., and Rabinovitch, M. Overexpression potentially recruited as pericyte-type cells. Pericyte coverage is of the serine elastase inhibitor elafin protects transgenic mice from hypoxic pulmonary hypertension. Circulation, 105: 516–521, 2002. known to stabilize vessel structures and protect against regression (33, 14. Ashcroft, G. S., Lei, K., Jin, W., Longenecker, G., Kulkarni, A. B., Greenwell-Wild, 34). Our observation that Wfdc1 (ps20) is a TGF-␤1 inducible gene T., Hale-Donze, H., McGrady, G., Song, X. Y., and Wahl, S. M. Secretory leukocyte relates directly to the potential vessel stabilizing effect of ps20, as the protease inhibitor mediates non-redundant functions necessary for normal wound healing. Nat. Med., 6: 1147–1153, 2000. direct interaction of endothelial cells with mural cells results in the 15. Wheeler, T. M., and Lebovitz, R. M. Fresh tissue harvest for research from prosta- activation of latent TGF-␤1 (35). Whereas a role for protease inhibi- tectomy specimens. Prostate, 25: 274–279, 1994. 16. Hirschi, K. K., Rohovsky, S. A., and D’Amore, P. A. PDGF, TGF-␤, and heterotypic tion in pericyte-induced vessel stabilization and maturation has yet to cell-cell interactions mediate endothelial cell-induced recruitment of 10T1/2 cells and be defined, the balance between protease activity and inhibition is their differentiation to a smooth muscle fate. J. Cell Biol., 141: 805–814, 1998. known to be central in the pathophysiology of large vessel destabili- 17. Gerdes, M. J., Dang, T. D., Lu, B., Larsen, M., McBride, L., and Rowley, D. R. Androgen-regulated proliferation and gene transcription in a prostate smooth muscle zation, as seen in abdominal aortic aneurysms (36, 37). cell line (PS-1). Endocrinology, 137: 864–872, 1996. Whereas ps20 is a small secreted protein of the WAP-type four- 18. Abruzzese, R. V., Godin, D., Mehta, V., Perrard, J. L., French, M., Nelson, W., disulfide core domain family, a target protease of ps20 inhibition has Howell, G., Coleman, M., O’Malley, B. W., and Nordstrom, J. L. Ligand-dependent regulation of vascular endothelial growth factor and erythropoietin expression by a not yet been identified. Although the WAP family member SLPI plasmid-based autoinducible GeneSwitch system. Mol. Ther., 2: 276–287, 2000. inhibits a broad range of target proteases, elafin is a specific inhibitor 19. Burcin, M. M., Schiedner, G., Kochanek, S., Tsai, S. Y., and O’Malley, B. W. Adenovirus-mediated regulable target gene expression in vivo. Proc. Natl. Acad. Sci. of elastase and proteinase 3 (11). For these reasons, we suspect that USA, 96: 355–360, 1999. ps20 inhibits a single protease or a narrow range of serine proteases. 20. Janik, P., Briand, P., and Hartmann, N. The effect of estrone-progesterone treatment Our laboratory is currently screening for a potential ps20 target on cell proliferation kinetics of hormone-dependent GR mouse mammary tumors. Cancer Res., 35: 3698–3704, 1975. protease. 21. Fox, S. B. Microscopic assessment of angiogenesis in tumors. In: J. C. Murray (ed.), Although additional studies are necessary to determine the exact Methods in Molecular Medicine: Angiogenesis Protocols, Vol. 46, pp. 29–46. To- mechanisms of action for ps20, this report indicates that ps20 is a towa, NJ: Humana Press Inc., 2001. 22. Jin, F. Y., Nathan, C., Radzioch, D., and Ding, A. Secretory leukocyte protease smooth muscle cell secreted protein capable of increasing endothelial inhibitor: a macrophage product induced by and antagonistic to bacterial lipopolysac- cell migration, and promoting tumor growth and angiogenesis. More- charide. Cell, 88: 417–426, 1997. 23. Gray, L. R., Alexander, A. L., and Shugars, D. C. Construction, non-denaturing over, the ps20 effects described in this report make the functional affinity purification, and characterization of baculovirally expressed human secretory inhibition of ps20 activity a possible target for antivascular tumor leukocyte protease inhibitor. Protein Expr. Purif, 26: 179–186, 2002. therapy. 24. Liekens, S., De Clercq, E., and Neyts, J. Angiogenesis: regulators and clinical applications. Biochem. Pharmacol., 61: 253–270, 2001. 25. Thompson, K., and Rabinovitch, M. Exogenous leukocyte and endogenous elastases ACKNOWLEDGMENTS can mediate mitogenic activity in pulmonary artery smooth muscle cells by release of extracellular-matrix bound basic fibroblast growth factor. J. Cell Physiol., 166: We thank Liz Hopkins for histological preparation of tissues. 495–505, 1996. 26. Tuxhorn, J. A., Ayala, G. E., and Rowley, D. R. Reactive stroma in prostate cancer progression. J. Urol., 166: 2472–2483, 2001. REFERENCES 27. Brawer, M. K., Deering, R. E., Brown, M., Preston, S. D., and Bigler, S. A. Predictors of pathologic stage in prostatic carcinoma. The role of neovascularity. Cancer (Phila.), 1. Tuxhorn, J. A., Ayala, G. E., Smith, M. J., Smith, V. C., Dang, T. D., and Rowley, 73: 678–687, 1994. D. R. Reactive stroma in human prostate cancer: induction of myofibroblast pheno- 28. Pedersen, H., Brunner, N., Francis, D., Osterlind, K., Ronne, E., Hansen, H. H., Dano, type and extracellular matrix remodeling. Clin. Cancer Res., 8: 2912–2923, 2002. K., and Grondahl-Hansen, J. Prognostic impact of urokinase, urokinase receptor, and 2. Tuxhorn, J. A., McAlhany, S. J., Dang, T. D., Ayala, G. E., and Rowley, D. R. type 1 plasminogen activator inhibitor in squamous and large cell lung cancer tissue. Stromal cells promote angiogenesis and growth of human prostate tumors in a Cancer Res., 54: 4671–4675, 1994. differential reactive stroma (DRS) xenograft model. Cancer Res., 62: 3298–3307, 29. Pedersen, H., Grondahl-Hansen, J., Francis, D., Osterlind, K., Hansen, H. H., Dano, 2002. K., and Brunner, N. Urokinase and plasminogen activator inhibitor type 1 in pulmo- 3. Tuxhorn, J. A., McAlhany, S. J., Yang, F., Dang, T. D., and Rowley, D. R. Inhibition nary adenocarcinoma. Cancer Res., 54: 120–123, 1994. of transforming growth factor-␤ activity decreases angiogenesis in a human prostate 30. Bajou, K., Masson, V., Gerard, R. D., Schmitt, P. M., Albert, V., Praus, M., Lund, cancer-reactive stroma xenograft model. Cancer Res., 62: 6021–6025, 2002. L. R., Frandsen, T. L., Brunner, N., Dano, K., Fusenig, N. E., Weidle, U., Carmeliet, 4. Larsen, M., Ressler, S. J., Lu, B., Gerdes, M. J., McBride, L., Dang, T. D., and G., Loskutoff, D., Collen, D., Carmeliet, P., Foidart, J. M., and Noel, A. The Rowley, D. R. Molecular cloning and expression of ps20 growth inhibitor. A novel plasminogen activator inhibitor PAI-1 controls in vivo tumor vascularization by WAP-type “four-disulfide core” domain protein expressed in smooth muscle. J. Biol. interaction with proteases, not vitronectin. Implications for antiangiogenic strategies. Chem., 273: 4574–4584, 1998. J. Cell Biol., 152: 777–784, 2001. 5. Rowley, D. R., Dang, T. D., Larsen, M., Gerdes, M. J., McBride, L., and Lu, B. 31. Bajou, K., Noel, A., Gerard, R. D., Masson, V., Brunner, N., Holst-Hansen, C., Purification of a novel protein (ps20) from urogenital sinus mesenchymal cells with Skobe, M., Fusenig, N. E., Carmeliet, P., Collen, D., and Foidart, J. M. Absence of growth inhibitory properties in vitro. J. Biol. Chem., 270: 22058–22065, 1995. host plasminogen activator inhibitor 1 prevents cancer invasion and vascularization. 6. Rowley, D. R. Characterization of a fetal urogenital sinus mesenchymal cell line U4F: Nat. Med., 4: 923–928, 1998. secretion of a negative growth regulatory activity. In Vitro Cell Dev. Biol., 28A: 32. Wen, W., Moses, M. A., Wiederschain, D., Arbiser, J. L., and Folkman, J. The 29–38, 1992. generation of endostatin is mediated by elastase. Cancer Res., 59: 6052–6056, 1999. 7. Larsen, M., Ressler, S. J., Gerdes, M. J., Lu, B., Byron, M., Lawrence, J. B., and 33. Gee, M. S., Procopio, W. N., Makonnen, S., Feldman, M. D., Yeilding, N. M., and Rowley, D. R. The WFDC1 gene encoding ps20 localizes to 16q24, a region of LOH Lee, W. M. Tumor vessel development and maturation impose limits on the effec- in multiple cancers. Mamm. Genome, 11: 767–773, 2000. tiveness of anti-vascular therapy. Am. J. Pathol., 162: 183–193, 2003. 8. Hennighausen, L. G., and Sippel, A. E. Mouse whey acidic protein is a novel member 34. Benjamin, L. E., Golijanin, D., Itin, A., Pode, D., and Keshet, E. Selective ablation of of the family of ’four-disulfide core’ proteins. Nucleic Acids Res., 10: 2677–2684, immature blood vessels in established human tumors follows vascular endothelial 1982. growth factor withdrawal. J. Clin. Investig., 103: 159–165, 1999. 9. Thompson, R. C., and Ohlsson, K. Isolation, properties, and complete amino acid 35. Hirschi, K. K., and D’Amore, P. A. Pericytes in the microvasculature. Cardiovasc. sequence of human secretory leukocyte protease inhibitor, a potent inhibitor of Res., 32: 687–698, 1996. leukocyte elastase. Proc. Natl. Acad. Sci. USA, 83: 6692–6696, 1986. 36. Sinha, S., and Frishman, W. H. Matrix metalloproteinases and abdominal aortic 10. Wiedow, O., Schroder, J. M., Gregory, H., Young, J. A., and Christophers, E. Elafin: aneurysms: a potential therapeutic target. J. Clin. Pharmacol., 38: 1077–1088, 1998. an elastase-specific inhibitor of human skin. Purification, characterization, and com- 37. Thompson, R. W., and Parks, W. C. Role of matrix metalloproteinases in abdominal plete amino acid sequence. J. Biol. Chem., 265: 14791–14795, 1990. aortic aneurysms. Ann. N. Y. Acad. Sci., 800: 157–174, 1996.

5865

Downloaded from cancerres.aacrjournals.org on October 5, 2021. © 2003 American Association for Cancer Research. Promotion of Angiogenesis by ps20 in the Differential Reactive Stroma Prostate Cancer Xenograft Model

Stephanie J. McAlhany, Steven J. Ressler, Melinda Larsen, et al.

Cancer Res 2003;63:5859-5865.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/63/18/5859

Cited articles This article cites 35 articles, 16 of which you can access for free at: http://cancerres.aacrjournals.org/content/63/18/5859.full#ref-list-1

Citing articles This article has been cited by 8 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/63/18/5859.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/63/18/5859. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.