Angiogenesis, Metastasis, and the Cellular Microenvironment Role of Plasminogen Activator Inhibitor-1 in Urokinase's Paradoxical In Vivo Tumor Suppressing or Promoting Effects

Tumor proteases and inhibitors have been associated with paradoxical effects on tumor progression in preclinical and clinical settings. We previously reported that urokinase (uPA) overexpression delays tumor progression in mammary cancer. This study aimed to determine the role of plasminogen activator inhibitor-1 (PAI-1) on uPA's paradoxical in vivo effects. Using syngeneic murine models, we found that stable uPA overexpression promoted in vivo growth of colon tumors (MC-38) naturally expressing high PAI-1, whereas growth inhibition was observed in renal tumors (RENCA) expressing lower PAI-1 levels. In murine mammary carcinoma (4T1), uPA overexpression shifted the uPA/PAI-1 balance in favor of the protease, resulting in significantly reduced tumor growth and metastases in vivo. Conversely, increased tumor progression was observed in stable PAI-1 overexpressing 4T1 tumors as compared with uPA-overexpressing and control tumors. These effects were associated with downregulation of metastases promoting genes in uPA-overexpressing tumors, such as metalloproteinases, CXCL-1, c-Fos, integrin a-5, VEGF-A, PDGF-a, and IL-1b. In PAI-1–overexpressing tumors, many of the above genes were upregulated. PAI-1 overexpressing tumors had increased total and new tumor microvessels, and increased tumor cell proliferation, whereas the opposite effects were found in uPA-overexpressing tumors. Finally, PAI-1 downregulation led to significant inhibition of 4T1 tumor growth and metastases in vivo. In conclusion, uPA's dual effects on tumor progression occur in the context of its interactions with endogenous PAI-1 expression. Our studies uncover novel mechanisms of in vivo tumor control by modulation of the balance between tumor proteases and inhibitors, which may be exploited therapeutically. Mol Cancer Res; 10(10); 1271–81. 2012 AACR. Introduction Tumor proteases have long been associated with tumor invasion, angiogenesis, and metastases (1, 2). It is widely accepted that urokinase (uPA), a member of the plasminogen activator (PA) system, is tumor promoting and associated with an aggressive tumor phenotype (1, 3). Tumor uPA expression is associated with shorter disease-free and overall survival in patients with early-stage breast cancer (4– 7) and has been proposed as a potential target for antitumor strategies (1, 3, 8–10). Paradoxically, overexpression of plasminogen activator inhibitor-1 (PAI-1), the endogenous inhibitor of uPA, is a clinically validated negative prognostic factor in breast and other cancers (6, 11, 12). PAI-1 has been shown to be essential for angiogenesis and tumor progression (13–15). PAI-1 promotes angiogenesis through interaction with vitronectin (16) and by direct inhibition of proteases (14). We have previously reported that PAs induce antiangiogenic effects in vitro and in vivo (17). We also showed that overexpression of uPA paradoxically delayed tumor growth, metastases, and improved survival in a syngeneic, immunocompetent mammary cancer model (18). uPA's tumordelaying effects were due to its protease activity, as tumors overexpressing proteolytically inactive uPA mutants were not associated with antitumor effects. On the other hand, several matrixmetalloproteinases (MMP) are associatedwith protective-–rather than promoting-–effects on in vivo tumor models (19). Proteases have been implicated in the generation of antiangiogenic peptides, such as angiostatin, endostatin, and tumstatin (20–22). The previous observations suggest that uPA can be both tumor promoting and protective and further support the notion that nonspecific inhibition of proteases may not necessarily prevent tumor progression (14, 19, 23). They also underscore the need to reevaluate current concepts on the role of uPA in cancer progression. The mechanisms of uPA-mediated tumor growth delay have not yet been characterized. In this report, we provide experimental evidence that the tumor promoting or delaying effects of uPA depend on its dynamic balance with tumor PAI-1.We also provide insight into the in vivo molecular changes that may mediate uPA's Authors' Affiliations: Sylvester Comprehensive Cancer Center; Department of Medicine; and Center for Computational Science, Miller School of Medicine, University of Miami, Miami, Florida Note: Supplementary data for this article are available at Molecular Cancer Research Online (http://mcr.aacrjournals.org/). Corresponding Author: Jaime R. Merchan, University of Miami Miller School of Medicine and Sylvester Comprehensive Cancer Center, 1475 NW12thAvenue, Suite 3400,Miami, FL 33136. Phone: 305-243-1287; Fax: 305-243-1293; E-mail: [email protected] doi: 10.1158/1541-7786.MCR-12-0145 2012 American Association for Cancer Research. Molecular Cancer Research www.aacrjournals.org 1271 tumor delaying effects, and show the antitumor and antimetastatic effects of in vivo PAI-1 inhibition. Materials and Methods Cell culture Murine mammary carcinoma 4T1 cell line, colon carcinoma MC38 cell line, renal cancer RENCA cell line, and 293T were obtained from American Type Culture Collection. Cells were grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% FBS at 37 C and 5% CO2. Generation of stable uPA and PAI-1–overexpressing cell lines cDNA-encoding murine uPA (obtained from ref. 18) and PAI-1 (gift from Foidart Jean-Michel, University of Liege, Wallonia, Belgium) were subcloned into the BamHI-NotI site of the lentiviral vector pHR-SIN-CSGWd1NotI (a gift of Y. Ikeda,Mayo Clinic, Rochester,MN), from pcDNA3.1 (þ)-muPA and pBS-mPAI-1, respectively, and the cDNA sequence was verified. Lentiviral packaging was conducted by cotransfection of the vector plasmidwith pCMV-Gag-Pol vector and pCMV-VSVG-poly-A vector into 293T cells using CaCl2 transfection kit (Promega). After 48 hours, lentivirus-containing supernatant was harvested and stored at 80 C. MC38 and RENCA cells were transduced with uPA-expressing lentiviral vector or empty vector (EV), respectively. Lentiviral constructs containing uPA, PAI-1, or an EV control were transduced into 4T1 cells. Approximately, 8 to 10 clones of each stable cell lines expressing uPA, PAI-1, or controls were isolated, and protein expression was assessed. Generation of stable PAI-1 knockdown 4T1 cell lines Lentiviral vectors (pGIPZ) containing short hairpin RNA (shRNA) against murine PAI-1 and nonsilencing controls were purchased fromOpenBiosystems. Lentiviral packaging and transduction into 4T1 cells were conducted following the manufacturer's recommendations. Ninety-six hours after transduction, stable PAI-1 shRNA–expressing cells were selected in 10 mg/mL puromycin (Sigma-Aldrich). PAI-1 knockdown was confirmed by real-time PCR (RTPCR) and determination of protein levels by ELISA. Western blot analysis Constructs for uPA wild-type and EV were transiently transfected into 293T cells as described earlier. After 48 hours, the maintenance culture medium was replaced with Opti-Medium. After 48 hours of continued incubation, the conditioned medium (CM) was collected and subsequently concentrated using the Centricon-10 filter units (Amicon). Protein concentrations were determined by using the BCA Protein Assay Kit (Pierce). Proteins were separated by SDSPAGE on 10% gels (Bio-Rad) under nonreducing and reducing conditions, transferred to nitrocellulose membrane (Amersham Biosciences), and immunoblotted with a monoclonal antibody against mouse uPA (sc-59727 Santa Cruz Biotechnology at 1:1,000 dilution). After the blots were washed by Tris-buffered saline with Tween (TBST), horseradish peroxidase–conjugated anti-mouse second antibody at 1:2,000 dilution (Cell signaling) was applied, and the peroxidase activity was revealed with the enhanced chemiluminescence system (Amersham Bioscience) according to the manufacturer's instructions. Quantification of uPA and PAI-1 Murine uPA and PAI-1 total protein and levels of active protein were determined in the CM (48 hours) of RENCA, MC38 and 4T1 clones by using a total or activemurine uPA, and PAI-1 ELISA kit, following the manufacturer's recommendations (Molecular Innovations). uPA proteolytic activity was assessed using a colorimetric urokinase activity kit (Chemicon), as previously reported (18). In vitro proliferation assay One thousand cells per well were plated in 96-well plates in CM and incubated at 37 C. Cell proliferation was determined by WST-1 proliferation reagent (Roche), as previously reported (18). Soft agar assay Soft agar assay was conducted using cell transformation detection kit (Millipore), as previously described (24). Briefly, 4T1 cells (2,500 cells/ well) were plated in DMEM plus 10% FBS in 0.4% agar on top of a 0.8% base agar layer in a 6-well plate. After 2 weeks, colony formation was quantified with a cell quantification solution (Millipore) for 3 hours at 37 C, followed by spectrophotometer reading at OD490. Cell migration and invasion Cell migration and invasion assays were separately conducted usingQCM cell migration and cell invasion (96 well) assay kits (Millipore), following the manufacturer's recommendations. Briefly, cells were harvested, suspended in serum-free medium, and plated at 5 10 cells per well onto the upper chamber in the presence of DMEM with 10% FBS in the bottom chamber. After 12 hours of incubation, the upper chamber was removed and placed in the additional 96-well tray containing cell detachment buffer with calcein-AM solution. Fluorescence in themigrated cells was measured at OD490. For cell invasion assay, 5 10 cells were plated in the top chamber, which had previously been coated withMatrigel. After 12 hours of incubation, the upper chamber was removed and fluorescence in the lower chamber was measured at OD490. Flow cytometry analysis Surface uPA receptor was detected by flow cytometry analysis, using a phycoerythrin-conjugated rat monoclonal anti-mouse uPAR (R&D systems), as previously described by us (25). Relative changes in cell surface uPAR expression levels were determined by quantitative assessment of fluorescence shifts (from flow-cytometry data) using WinMDI 2.9 software (J. Trotter, Scripps Research Institute, La Jolla, CA) and expressed as fold changes of the mean fluorescence index, as described (25). Jing et al. Mol Cancer Res; 10(10) October 2012 Molecular Cancer Research 1272 In vivo studies Animal studies were approved by Institutional Animal Care and Use Committee of University of Miami and the Mayo Clinic. MC38 and RENCA (5 10 in 50 mL PBS) cells were inoculated into the flank of C57BL/6 (n 1⁄4 6–7 per group) and BALB/C (n 1⁄4 8–10/group) mice, respectively. 4T1 cells (10 cells) were implanted into the fifth mammary fat pad of 8to 10-week-old female BALB/c mice (5 mice/group), as described (18). Tumor volume was measured every 3 days and calculated using the standard formula (width length 0.52). In the experiments involving 4T1 tumors, mice were followed until day 34 when they were sacrificed. Lungs were removed and fixed in Bouin's solution, and surface lung metastases were analyzed (26). In an additional experiment, 4T1 cells (10 cells) were implanted into BALB/c mice (12 mice/group). Tumor volume was measured as above until day 21, when tumors were resected and mice continued to be followed. At day 42, they were sacrificed and lungs were resected (randomly) from 5 mice per group. Lungs were fixed (10% neutral-buffered formalin), embedded in paraffin, sectioned (5-mm slices, 400 mm apart), and stained with hematoxylin and eosin. Magnified ( 20) lung nodules were counted. Isolation of tumor cells from primary tumor tissues uPA-overexpressing 4T1 cells were stably transfected with a lentiviral vector expressing GFP (pHR-SIN-GFP). GFPpositive 4T1 cells were implanted BALB/c mice as earlier (n1⁄4 3). At day 7 after inoculation, freshly isolated tumorwas treated with 10 mg/mL collagenase I (Sigma). The cell suspension was filtered through 70-mm nylon mesh and washed in Hank's balanced salt solution. GFP-positive and -negative cells were separated by fluorescence-activated cell sorting (FACS; BD FACS Aria-I). RNA purification and gene expression array Total RNA was extracted from tumor samples (triplicate) using Qiagen total RNA isolation kit. The Illumina MouseWG-6 v2.0 Expression BeadChip (Illumina) was used for gene expression. The microarray data have been deposited in National Center for Biotechnology Information's (NCBI) Gene Expression Omnibus (GEO, accession number GSE38346). The raw data from the fluorescence intensity measurements of each array experiment were processed using GeneSpringGX v.11.0 software (Agilent). Statistical analysis, fold-change calculations, and hierarchical clustering of the data were also conducted in GeneSpring software. Genes that expressed significantly differently with more than 1.5fold change and a P value of <0.05 with respect to controls were taken into consideration. Gene expression data were further validated by quantitative real-time (RT)-PCR (qRTPCR) analysis. Pathway analysis was conducted by MetaCore software (GeneGo, Inc.). Real-time PCR analysis The assay was conducted from RNA tumor samples (n1⁄4 3 per group) using TaqMan one-step PCRmaster mix reagents kit and Applied Biosystems 7300 qPCR system. Premade primers and probes were purchased fromApplied Biosystems: glyceraldehyde-3-phosphate dehydrogenase (GAPDH), MMP-9, MMP-10, MMP-12, MMP-13, CXCL1, JAG-1, endothelin 1 (Edn1), ADAMTS12, c-Fos, PDGFa, integrin a-5 (Itga5), VEGF-A, and IL-1 b. Each measurement was carried out in triplicate. Differences in gene expression, expressed as fold change, were calculated using the 2 DDCt method using GAPDH as the internal control. Immunohistochemistry/immunofluorescence studies Tumor samples were frozen in OCT compound and sectioned (8 mm). Slides were fixed with cold acetone for 5 minutes. Total and neovessels were detected with Alexa Fluor 488 conjugated lectin (1:1,000 dilution; Invitrogen) and Alexa Fluor 488 conjugated CD105 (Abcam), respectively, as previously described (27, 28). Cell nuclei were stained with 40,6-diamidino-2-phenylindole (DAPI; 1:5,000; Invitrogen). Tumor microvessels were assessed with a photomicroscope (Nikon, NE2000). Pictures were taken at 400 magnification. Tumor cell proliferation was assessed by Ki-67 staining as follows: cryostat sections were fixed in cold acetone for 10 minutes, and endogenous peroxidase activity was quenched with 0.3% H2O2 for 10 minutes at room temperature. The slides were washed in PBS and incubated with rabbit anti-Ki67 antibody (Millipore) for 30 minutes at 37 C. After washing in PBS, the slides were developed with VECTASTAIN ABC (avidin–biotin complex) peroxidase kit (Vector Laboratories) and a 3,3,9-diaminobenzidine (DAB) peroxidase substrate kit (Vector Laboratories) according to the manufacturer's instructions. Statistical analysis Data are presented as mean SD or SEM. In vitro experiments were carried out in triplicate and repeated twice, unless otherwise specified. Differences in mean from the in vitro experiments were compared using the Student t test and Wilcoxon rank-sum test. Differences were considered statistically significant at P < 0.05. Differences of in vivo tumor growth and lung metastases among 3 or more groups were analyzed by one-way ANOVA. Pairwise comparisons were conducted using the Tukey–Kramer method. All statistical tests were 2-sided. Results Urokinase overexpression is associated with tumor promoting or suppressing effects in different syngeneic models: role of endogenous PAI-1 expression We have previously reported that uPA overexpression delays in vivo tumor progression in a syngeneic mammary cancer model (18). To extend the previous findings, we generated stably uPA-overexpressing murine renal cancer (RENCA-uPA) and murine colon cancer (MC-38-uPA) cell lines using lentiviral vector technology, as described in Materials and Methods. Transduction of cells with the lentiviral vector induces expression of high molecular weight Changes in uPA/PAI-1 Balance Regulate Tumorigenesis www.aacrjournals.org Mol Cancer Res; 10(10) October 2012 1273 uPA, with a migration pattern similar to active 2-chain mouse high molecular weight (HMW) uPA used as positive control (Supplementary Fig. S1A and S1B). Significant uPA overexpression was confirmed in the CM from stable RENCA (Fig. 1A) and MC-38 (Fig. 1B) cells as compared with controls. Levels of active uPA were increased in RENCA-uPA cells (Fig. 1A); however, in the CM of MC-38-uPA cells, levels of active uPA were significantly lower than their RENCA counterparts (P< 0.0001; Fig. 1B). This was associated with higher uPA proteolytic activity (chromogenic assay) in the RENCA uPA CM as compared with MC-38-uPA CM (Supplementary Fig. S1C and S1E). 15