Cdna Microarray Analysis of Pigment Epithelium-Derived Factor-Regulated

International Journal of Urology (2009) 16, 323–328 doi: 10.1111/j.1442-2042.2008.02199.x

Original Article: Laboratory Investigation cDNA microarray analysis of pigment epithelium-derived factor-regulated gene expression proﬁle in prostate carcinoma cells Weiwei Liu,1 Zhong Wu,2 Ming Guan1 and Yuan Lu1 Department of 1Laboratory Medicine and 2Urology, Huashan Hospital, Shanghai Medical College, Fudan University, Shanghai, China

Objectives: To clarify molecular mechanisms involved in the action of pigment epithelium-derived factor (PEDF) in hormone insensitive prostate cancer cells. Methods: Total ribonucleic acid from untreated and PEDF-treated cells was subjected to microarray analysis using BioStar 8464 microarray. Real-time polymerase chain reaction analysis was conducted to confirm the microarray data. Results: Twenty-seven out of 8464 genes were found altered in both cell lines. Common gene responses altered by PEDF were identified and included genes known to alter cell signaling as well as genes involved in catalytic activity, cell proliferation, angiogenesis and apoptosis. Real-time reverse transcription polymerase chain reaction, in accordance with the microarray analysis, indicated that PEDF treatment caused an upregulation in the mRNA expression level of stanniocalcin 2, brain-specific angiogenesis inhibitor 2 and growth arrest, DNA-damage-inducible, alpha, and downregulation in the messenger ribonucleic acid level of fibroblast growth factor 3, teratocarcinoma-derived growth factor, neuropilin1, and endothelial Per/ARNT/Sim domain protein1, respectively. Conclusions: These findings demonstrate that PEDF administration causes significant changes in the gene expression of the prostate, providing insights into the potential role of PEDF in the treatment of prostate cancer.

Key words: angiogenesis, microarray, pigment epithelium-derived factor, prostate cancer.

Introduction clearly elucidated. Here, for the first time, we utilized a high- throughput microarray, which contains 8464 known genes, to deter- Pigment epithelium-derived factor (PEDF), a multifunctional secre- mine the alternation of gene expression profiles of hormone tory soluble glycoprotein that belongs to the superfamily of serine insensitive cell DU145 and PC-3 exposed to PEDF. protease inhibitors, is the most potent endogenous inhibitor of angiogenesis in various tissues. PEDF is reported to inhibit retinal endothelial cell growth, migration and to suppress ischemia-induced retinal Methods 1 neovascularization. Ocular administration of an adenoviral construct Cell lines containing the PEDF gene inhibits the formation of ocular angiogenesis.2 Furthermore, loss of PEDF is associated with metastatic sub- Cell lines established from prostate cancer (DU145, PC-3) were clone of some tumors and there is allelic loss of PEDF genes in obtained from American Type Culture Collection (Rockville, MD). others.3 Ectopic PEDF expression delays the growth and invasion of PC-3 cells were cultured in RPMI 1640 (Invitrogen, San Diego, CA) lung carcinoma, hepatocellular carcinoma, melanoma, and glioblas- with 10% fetal bovine serum (FBS) (Sigma, St. Louis, MO). DU145 toma, where it blocks neovascularization.4–9 One clinical observation cells were maintained in DMEM with 10% FBS. The 293T cells were indicates that loss of PEDF expression may be associated with pro- obtained from ATCC (Rockville, MD) and maintained in Dulbecco’s gression toward a metastatic phenotype in prostate cancer.10 PEDF can modified Eagle’s medium (Invitrogen) containing 10% FBS. The inhibit stromal vasculature and epithelial growth in the xenograft immortalized normal prostate epithelial cell lines RWPE-1 (American tumor model of prostate cancer. In PEDF-deficient mice, microvascu- Type Culture Collection) were cultured in keratinocyte serum-free lar density was increased and associated with hyperplasia. PEDF is medium supplemented with 5 ng/mL human recombinant epidermal suppressed by androgen in the cultured prostate epithelium and growth factor and 0.05 mg/mL bovine pituitary extract (Invitrogen, increased in the prostate in vivo upon castration, indicating that PEDF Carlsbad, CA). is an angiogenesis inhibitor in this organ.11 Moreover, the PEDF pep- tides owned distinct anticancer activities on PC-3 tumor growth. The 3-(4, 5-Dimethylthiazolyl-2)-2,5-Diphenyltetrazolium 34-mer (residues 24–57) decreased tumor microvessel density and bromide assay increased apoptosis, whereas the 44-mer (residues 78–94) prompted neuroendocrine differentiation of PC-3 cells. However, the underlying Cells grown in 96-well tissue culture plates were treated with various mechanisms of PEDF that interfere with tumor growth have not been doses of PEDF for 48 h and incubated with the reagent in the Cell Titer 96 Aqueous One solution cell proliferation assay kit (Promega, Correspondence: Professor Yuan Lu, Department of Laboratory Medicine, Madison, WI). The absorbance of reduced tetrazolium compound Huashan Hospital, 12 Central Urumqi Road , Shanghai, 200040, China. Email: derived from the reagent due to dehydrogenase activities in viable cells [email protected] was recorded with a microtiter plate reader (Molecular Devices Co. Received 6 March 2008; accepted 1 October 2008. Sunnyvale, CA) at a test wavelength of 450 nm and a reference wave- Online publication 12 January 2009 length of 650 nm.

Construction of the complementary DNA for Cy3). Genes were identiﬁed as differentially expressed if the ratio of expression of N-terminally hexahistidine-tagged (Cy5/Cy3) ¥ normalization cofactor was >1.50 or <0.67. PEDF (N-His-PEDF)

The primers for PEDF were designed, with an EcoRI overhang on the Real-time reverse transcription polymerase 5′ end and an XbaI overhang on the 3′ end. The sequences were as chain reaction follows: forward strand, 5′-atgaattcatgcaggccctggtgcta-3′; reverse Based on the results from the arrays, several genes that showed alter- strand, 5′-attctagattaggggcccctggggtcca-3′. The PEDF complementary ations in expression after exposure to PEDF were chosen for more DNA (cDNA) was inserted into the EcoRI /XbaI site of the mammalian thorough investigation. First strand cDNA was synthesized from total expression vector pcDNA3.1/HIS C (Invitrogen). The 293T cells were RNA (2 mg) each using Superscript II reverse transcriptase (Invitrogen) grown in Dulbecco’s modified Eagle’s medium supplemented with and OligdT primers. The primers for stanniocalcin 2(STC2), 10% fetal calf serum, and incubated at 37°C in the presence of 5% CO . 2 guanine nucleotide binding protein (G protein), alpha 13 (GNA13), The 293T cells were grown to confluence in T150 flasks and transiently teratocarcinoma-derived growth factor 1 (TDGF1), fibroblast growth transfected with 60 mg of vector pcDNA3.1/HIS-PEDF and 120 ml factor 3 (FGF3), brain-specific angiogenesis inhibitor 2 (BAI2), FuGENE 6 transfection reagent (Roche Diagnostics, Mannheim, neuropilin 1 (NRP1), endothelial Per/ARNT/Sim (PAS) domain Germany) per flask. Transient transfections were performed as protein 1 (EPAS1), growth arrest and DNA-damage-inducible, alpha described by the manufacturer. Hexahistidine-tagged PEDF proteins (GADD45A) from RT2 Primer Set (SuperArray, Frederick, MD) were were purified from conditioned media by a Ni-NTA spin kit (Qiagen used in real-time PCR reactions with SYBR Green. Real-time PCR GmbH, Hilden, Germany) according to the manufacturer’s instruc- (25 ml reaction volume) was performed with the appropriate primers tions. (SuperArray) as per the manufacturer’s instructions in triplicate with PCR Master Mix (SuperArray) at 95°C for 15 min and amplified for 40 cycles (95°C, 30 s, 55°C, 30 s, 72°C, 30 s) and primer pairs for 18S Western blot analysis ribosomal RNA as reference RNA. Cycle time (Ct) was measured using After determining the protein concentration by means of the Bio-Rad the ABI 7000 (Applied Biosystem, Foster City, CA). In all experiments, protein assay, an equal amount of purified PEDF protein (20 mg) from and mean threshold cycle (Ct) values were calculated. Quantification of each sample was subjected to SDS–PAGE (10%) and transferred onto a given gene, expressed as fold induction over control (vehicle-treated), CT a polyvinylidene difluoride membrane. The membrane was then treated was calculated after normalization to GAPDH using the 2-DD . with mouse monoclonal antibody against human PEDF (Chemicon, Temecula, CA). The resultant immunocomplexes were visualized with Results an enhanced chemiluminescence system according to manufacturer’s instructions. Preparation of recombinant PEDF

The PEDF proteins expressed using pcDNA3.1/HIS C possess a His Microarray analysis Tag at their N-terminus and therefore can be purified easily from conditioned media where the cells were grown, using Ni-NTA column The cDNA microarray containing a set of 8464 sequence verified chromatography. As shown in Figure 1, Western blot analysis of puri- human cDNA clones was provided (Biostar Genechip Inc., Shanghai, fied PEDF proteins revealed a single band with a molecular weight of China). The microarray hybridization and data analysis were accom- about 50 kDa, which showed positive reactivity with mouse mono- plished as follows: Total RNA was extracted from cells using the clonal antibody against human PEDF. Under normal culture conditions, RNeasy system (Qiagen). Messenger RNA converted to cDNA and the 0.33 mg of recombinant protein was purified from 1 ml of cell culture probes from vehicle-treated (DMSO) cells were labeled with cyanine supernatant. 3-dUTP (Cy3), those from cells treated with 100 nmol/L PEDF were labeled with cyanine 5-dUTP (Cy5), respectively, using the LabelStar Cell Growth Inhibition by PEDF system (Qiagen). Probe mixtures were precipitated by ethanol and suspended in 20 ml of hybridization buffer (5¥ SSC + 0.2% SDS). The DU145 and PC-3 prostate cancer cells and normal prostate epithelial purified cDNA was resuspended in 80 ml of hybridization solution RWPE-1 cells were treated with DMSO (control), 1, 10 and containing 5¥ SSC, 0.1% SDS, 50% formamide, 20 mg of rat Cot-1 DNA, 20 mg of polyadenylic acid potassium salt (Sigma) and 20 mgof yeast tRNA (Invitrogen, Carlsbad, CA). The hybridization mixtures were heated at 100°C for 2–3 min and directly pipetted onto microarrays. The arrays were then hybridized at 42°C for 12–16 h in a humidity hybridization chamber. The hybridized microarrays were washed in 2¥ SSC + 0.1% SDS for 5 min, 0.1¥ SSC + 0.1% SDS for 5 min at room temperature. After a quick rinse in 0.1¥ SSC, the slides were air-dried for scanning. Primary data were processed using Gene Pix4000B (Axon Instruments) to determine the intensities of each spot at the two wavelengths representing the quantity of Cy3-dUTP and Cy5-dUTP, respectively. To minimize artifacts arising from low expression, only Fig. 1 Western blot analysis of conditioned medium from 293T cells, tran- genes whose Cy3 and Cy5 fluorescent intensities were both over 200 siently transfected with the expression vector alone (lane 1) or with human counts, or genes whose either Cy3 or Cy5 fluorescent intensity was over pigment epithelium-derived factor expression vector (lane 2) or human 500 were selected for calculating the normalization cofactor (Cy5/ vitreous humor is as positive control (lane 3).

100 nmol/L PEDF over 2 days. There was no morphological change gene transcripts differentially regulated following treatment with between vehicle-treated (DMSO) cells and cells treated with a series of PEDF. Two independent cDNA microarray analyses were used to levels of PEDF. The effect of PEDF on the proliferation of PC-3 and compare gene expression patterns between two populations of each cell DU145 cells is depicted in Figure 2. A significant inhibition in cell line, one of which was treated for 48 h with 100 nM PEDF while the viability was observed only upon treatments with 100 nmol/L and cell other was treated as vehicle (DMSO). The signal differences for most viability was inhibited by 24.6% and 33.8% at 48 h in DU145 and of the expressed genes between the control and PEDF-treated mRNA PC-3, respectively, compared with control cells. However, no signifi- populations were within twofold. Only those genes that exhibited a fold cant effect on cell viability was observed in RWPE-1 cells. The con- induction or reduction of 1.5 or greater from duplicate experiments in centration of 100 nmol/L PEDF, which show marked reduction in both DU145 and PC-3 cells were selected as differentially expressed. In cancer cell viability, was used for the microarray analysis. a comparison of two independently prepared and hybridized samples, the hybridizations showed linear slope in each group (Fig. 3a,b). High Microarray analysis levels of correlation from two independent samples of the same group reflect the accuracy of gene expression data derived from the DNA To better understand the mechanism of the action of PEDF on the arrays. prostate cells, the study was performed using microarrays to identify The selection criteria adopted to identify differentially expressed genes yielded seven genes whose expression is stimulated and 20 genes 110 DU145-PEDF(100nM) whose expression is suppressed by PEDF (Table 1). Among the differ- PC-3-PEDF(100nM) entially expressed, mRNA encoding proteins with signal transduction RWPE-1-PEDF(100nM) constituted the largest group (n = 11) followed by catalytic activity 100 (n = 4), proliferation (n = 3), angiogenesis (n = 3), protein binding (n = 2), transporter activity (n = 2), apoptosis (n = 1), and neurogenesis 90 (n = 1).

80 Real-time PCR

Viability (% of Control) 70 Notably, the most important downstream mediators after treatment with 60 PEDF were STC2, GNA13, TDGF1, FGF3, BAI2, NRP1, EPAS1 and 1nM 10nM 100nM GADD45A. These eight differentially regulated genes were selected and examined by real-time RT–PCR. These genes were chosen based Fig. 2 Cell viability of DU145, PC-3 and RWPE-1 under the treatment of on their potential roles in tumor proliferation and our interest for pigment epithelium-derived factor (PEDF). Cells were treated for 48 h with further studies. As shown in Figure 4, the Quantitative reverse tran- different concentrations of PEDF and viability was assessed by MTT assay. scription polymerase chain reaction (RT–PCR) measurements in Results are expressed as percent of corresponding control and represen- general correlated positively with the microarray data except GNA13, tative of three duplicated independent experiments. which were detected by microarray analysis, and were not identiﬁed by

AB 50000 50000 CY3 Signal

10000 10000 0 30000 0 30000

0 10000 20000 30000 40000 50000 60000 0 10000 20000 30000 40000 50000 60000

CY5 Signal CY5 Signal

Fig. 3 Scatter plot of expressed transcripts from DU145 (a) and PC-3(b). The hybridization signal intensities from Cy5 signal (pigment epithelium-derived factor treatment) were plotted against the intensities of Cy3 signal (Control). Replicate experiments exhibited similar scatter plots.

Table 1 Summary of genes differentially expressed in DU145 and PC-3 after treatment with pigment epithelium-derived factor

Functional classiﬁcation UniGene Gene name Fold change in Microarray

DU145 PC-3

Signal transducer Hs.446678 nuclear receptor coactivator 2 1.73 2.19 Hs.438482 WD repeat domain 19 (WDR19) 2.59 2.14 Hs.233160 stanniocalcin 2 1.64 2.11 Hs.50716 signal-regulatory protein beta 2 0.66 0.64 Hs.515018 guanine nucleotide binding protein (G protein), alpha 13 0.650 0.64 Hs.469386 inositol polyphosphate-4-phosphatase, type I 0.63 0.65 Hs.303719 protein kinase, cGMP-dependent, type II 0.62 0.64 Hs.406156 dedicator of cytokinesis 7 0.66 0.58 Hs.462457 A kinase (PRKA) anchor protein 10 (AKAP10), nuclear gene encoding 0.65 0.58 mitochondrial protein Hs.43322 protein kinase, AMP-activated , alpha 1 catalytic subunit 0.59 0.49 Hs.506852 protein tyrosine phosphatase, non-receptor type 11 (Noonan syndrome 1) 0.43 0.65

Catalytic activity Hs.448851 ubiquitin speciﬁc protease 6 0.64 0.65 Hs.309958 guanylate cyclase 2D, membrane (retina-speciﬁc) 0.58 0.63 Hs.499793 phosphatidylinositol glycan, class L 0.62 0.57 Hs.477 hydroxysteroid (17-beta) dehydrogenase 3 0.58 0.57

proliferation Hs.34012 breast cancer 2, early onset 0.62 0.61 Hs.385870 teratocarcinoma-derived growth factor 1 0.62 0.61 Hs.37092 ﬁbroblast growth factor 3 0.60 0.46

Angiogenesis Hs.524138 brain-speciﬁc angiogenesis inhibitor 2 1.61 1.82 Hs.131704 neuropilin 1 0.48 0.51 Hs.468410 endothelial PAS domain protein 1 0.44 0.49

Protein Binding Hs.518460 adaptor-related protein complex 2, mu 1 subunit 1.62 1.64 Hs.28608 zinc ﬁnger, CCHC domain containing 3 0.58 0.60

Transporters Hs.302738 solute carrier family 26 (sulfate transporter), member 2 0.63 0.57 Hs.282151 calcium channel, voltage-dependent, alpha 2/delta subunit 1 0.56 0.50

Apoptosis Hs.80409 growth arrest and DNA-damage-inducible, alpha 2.5 3.3

Neurogenesis Hs.184945 gastrulation brain homeobox 2 1.52 1.87

AMP, adenosine monophosphate; CCHC, Cys-Cys-His-Cys; PAS, Per/ARNT/Sim; PRKA, adenosine 5′ monophosphate-activated protein kinase.

real-time RT–PCR. This may be due at least partly to differences in mulating evidence that PEDF, a member of the serpin superfamily of methodology. Quantitative PCR was performed to confirm that these protease inhibitors, is associated with cell differentiation and angiogen- mRNAs were indeed differentially regulated. esis, more attention should be focused on potential antitumor effect. However, the mechanisms by which PEDF exhibits antitumor effect on Discussion solid tumors are poorly defined. In this study we have shown that signal transducer genes, cell proliferation genes and angiogenesis genes were Progression to androgen independence remains the main obstacle to differentially displayed in PEDF-induced prostate cancer cells. These improving survival for patients with advanced prostate cancer (PCa). novel findings have important implications for improving the efficacy Despite the general success of anti-androgen therapy, an androgen- of chemotherapy of prostate cancer and for developing new chemo- refractory status almost invariably develops with time, with an eventu- therapeutic drugs. ally fatal outcome.12,13 Because tumor spreading is responsible for the Stanniocalcin 2 (STC2) is homodimeric glycoprotein that is majority of deaths of cancer patients, the development of therapeutic expressed in a wide variety of tissues and plays a role in phosphate agents that inhibit tumor metastasis is very desirable. Given the accu- transport and calcium concentration.14 High STC2 mRNA levels are

A B

3 3

2 2

1 1

TDGF1 FGF3 NRP1 EPAS1 GNA13 TDGF1 FGF3 NRP1 EPAS1 GNA13 0 0 STC2 BAI2 GADD45A STC2 BAI2 GADD45A Fold change Fold change -1 -1

-2 -2

-3 -3

-4

Fig. 4 Real-time reverse transcription polymerase chain reaction (RT–PCR) of relative expression of STC2, BAI2, GADD45A, TDGF1, FGF3, NRP1, EPAS1 and GNA13 in pigment epithelium-derived factor (PEDF)-treated DU145 cell (A) and PC-3 cell (B) compared with corresponding control (n = 3). Columns, mean messenger ribonucleic acid (mRNA) expression of each gene; bars, standard error. Upregulation of STC2, BAI2 and GADD45A as well as downregulation of TDGF1, FGF3, NRP1and EPAS1were found in cell exposed to 100 nmol/L PEDF. GNA13 was not identified by real time RT–PCR. significantly associated with good prognosis in hormone receptor- BAI1. The seven-span transmembrane region (STR) and two functional positive breast cancer patients.15 Upregulation of STC2 may act as a elements, an Arg-Gly-Asp (RGD) motif and thrombospondin type 1 potent growth inhibitor as it has been shown in transgenic mice.16 repeats (TSR) are well conserved between BAI1 and BAI2. During We identified two oncogenic (oncofetal) genes that are expressed at postischemic angiogenesis in the rat focal cerebral ischemia injury lower levels in PEDF treated cells than in vesicle-treated cells. The model, vascular endothelial growth factor (VEGF) expression is teratocarcinoma-derived growth factor-1 (TDGF-1) gene and fibroblast increased following the decreased BAI2 expression. Upregulation of growth factor (FGF-3). TDGF-1, a member of the epidermal growth VEGF was observed in the antisense BAI2 cDNA-transfected cells and factor (EGF)–Cripto/FRL-1/Cryptic (CFC) family, has been implicated indicated the inhibitory effect of angiostatic BAI2 on angiogenic VEGF in embryogenesis and in carcinogenesis. TDGF-1 possesses an EGF- expression.26 Neuropilin-1(NRP-1), a VEGF165 isoform-specific like consensus sequence that contains six cysteine residues in a region receptor, has been identified by Soker et al.27 Expression of NRP-1 has of approximately 37 amino acids.17 In EpH-4 mammary epithelial cells, recently been found in breast cancer, melanoma and colon cancer.28–30 TDGF-1 can induce the specific phosphorylation of Smad-2 in the Overexpression of NRP-1 in rat prostate carcinoma cells results in presence of Nodal and induce the activation of both the ras/raf/MAPK increased tumor growth in vivo as well as increased microvessel density and PI3K/Akt pathways in mammalian cells.18,19 The exact role of and endothelial cell proliferation.31 Endothelial PAS domain protein TDGF-1 in the development of PCa is not clear and needs to be 1(EPAS1) was recognized as key transcription factors regulating the determined. FGF-3 belongs to the FGF gene family and its inductive expression of a variety of genes involved in angiogenesis, glycolysis and role in cell proliferation and migration has been postulated; this gene apoptosis.32 EPAS1 overexpression by tumor cells may therefore be could contribute to the development of some human tumors. By devel- endogenous markers of hypoxia and it stimulates angiogenesis through oping biogenic mice that express the FGF-3 target transgene under the activation of the VEGF gene. Most recently, Acker et al.33 reported that prostate, Chua et al. found ectopic FGF-3 expression severely perturbs overexpression of EPAS1 enhanced glioma tumor vascularization but normal prostate development and increases in epithelial stratification impaired tumor growth, in part because of an increase in tumor cell and basophilia similar to prostatic intraepithelial neoplasia (PIN)-like apoptosis. Therefore, PEDF could inhibit tumor angiogenesis, not only lesions, suggesting its role in proliferation in the prostate.20 through the well-known anti-angiogenic effects on endothelial cells,22 It is well known that PEDF induces apoptosis in some cancer but also through upregulation of BAI 2 and downregulation of NRP-1 cells.21,22 In this study, we found that PEDF also regulated the expres- and EPAS1 in prostate carcinoma cells. sion of GADD45A, which is critically involved in the apoptotic pro- These results may represent intrinsic cellular response to this agent cesses. GADD45A gene expression is regulated by the TP53 protein.23 as both anti-cancer and anti-angiogenesis mechanisms. We have Increased expression of GADD45a protein arrests the cell cycle at the recently noted that forced overexpression of PEDF has direct growth G2/M phase.24 The induction of apoptosis mediated by this gene could inhibitory action on the prostate carcinoma cell and glioma cell through be another molecular mechanism by which PEDF inhibits the growth of induction of tumor cell apoptosis and cell arrest in vitro and in vivo,in prostate cancer cells. which PEDF may control these differentially expressed genes.34,35 Nev- It was shown that PEDF inhibited angiogenic processes and was more ertheless, the findings in this study are initial and limited to a small set effective than the well studied angiogenesis inhibitor angiostatin.25 of cell lines. Further studies including large-scale clinical tumor speci- Three pro- or anti-angiogenic factors capable of modulating vessel men and in vivo transcriptional analysis should be undertaken. growth were differently induced by PEDF treatment. Brain-specific In conclusion, PEDF treatment directly and indirectly caused angiogenesis inhibitor 2 (BAI2) is a novel human gene homologous to changes in the expression of many genes that are critically involved in

© 2009 The Japanese Urological Association 327 W LIU ET AL. the control of cell proliferation, apoptosis, oncogenesis, and angiogen- 17 Saloman DS, Bianco C, Ebert AD et al. The EGF-CFC family: Novel esis. PEDF could be a promising agent for the prevention and or epidermal growth factor-related proteins in development and cancer. treatment of prostate cancer and its metastasis. Endocr. Relat. Cancer 2000; 7: 199–226. 18 Kannan S, De Santis M, Lohmeyer M et al. Cripto enhances the tyrosine phosphorylation of Shc and activates mitogen-activated protein Acknowledgments kinase (MAPK) in mammary epithelial cells. J. Biol. Chem. 1997; 272: 3330–5. This work was supported in part by Natural Science Foundation of 19 Ebert AD, Wechselberger C, Nees M et al. Cripto-1-induced increase in China (No.30872590) and the Foundation of Shanghai Medical Key vimentin expression is associated with enhanced migration of human Discipline. Caski cervical carcinoma cells. Exp. Cell Res. 2000; 257: 223–9. 20 Chua SS, Ma ZQ, Gong L, Lin SH, DeMayo FJ, Tsai SY. Ectopic References expression of FGF-3 results in abnormal prostate and Wolffian duct development. Oncogene 2002; 21: 1899–908. 1 Duh EJ, Yang HS, Suzuma I et al. Pigment epithelium-derived factor 21 Stellmach V, Crawford SE, Zhou W, Bouck N. Prevention of suppresses ischemia-induced retinal neovascularization and ischemia-induced retinopathy by the natural ocular antiangiogenic agent VEGF-induced migration and growth. Invest. Ophthalmol. Vis. Sci. pigment epithelium-derived factor. Proc. Natl. Acad. Sci. U. S. A. 2001; 2002; 43: 821–9. 98: 2593–7. 2 Gehlbach P, Demetriades AM, Yamamoto S et al. Periocular injection of 22 Takenaka K, Yamagishi S, Jinnouchi Y, Nakamura K, Matsui T, an adenoviral vector encoding pigment epithelium-derived factor inhibits Imaizumi T. Pigment epithelium-derived factor (PEDF)-induced choroidal neovascularization. Gene Ther. 2003; 10: 637–46. apoptosis and inhibition of vascular endothelial growth factor (VEGF) 3 Guan M, Yam HF, Su B et al. Loss of pigment epithelium-derived expression in MG63 human osteosarcoma cells. Life Sci. 2005; 77: factor expression in glioma progression. J. Clin. Pathol. 2003; 56: 3231–41. 277–82. 23 Smith ML, Ford JM, Hollander MC et al. p53-mediated DNA repair 4 Abe R, Shimizu T, Yamagishi S et al. Overexpression of pigment responses to UV radiation: Studies of mouse cells lacking p53, p21, epithelium-derived factor decreases angiogenesis and inhibits the growth and/or gadd45 genes. Mol. Cell Biol. 2000; 20: 3705–14. of human malignant melanoma cells in vivo. Am. J. Pathol. 2004; 164: 24 Maeda T, Hanna AN, Sim AB, Chua PP, Chong MT, Tron VA. 1225–32. GADD45 regulates G2/M arrest, DNA repair, and cell death in 5 Mahtabifard A, Merritt RE, Yamada RE, Crystal RG, Korst RJ. In vivo keratinocytes following ultraviolet exposure. J. Invest. Dermatol. 2002; gene transfer of pigment epithelium-derived factor inhibits tumor 119: 22–6. growth in syngeneic murine models of thoracic malignancies. J. Thorac. 25 Dawson DW, Volpert OV, Gillis P et al. Pigment epithelium-derived Cardiovasc. Surg. 2003; 126: 28–38. factor: A potent inhibitor of angiogenesis. Science 1999; 285: 245–8. 6 Wang L, Schmitz V, Perez-Mediavilla A, Izal I, Prieto J, Qian C. 26 Kee HJ, Koh JT, Kim MY et al. Expression of brain-specific Suppression of angiogenesis and tumor growth by adenoviral-mediated angiogenesis inhibitor 2 (BAI2) in normal and ischemic brain: gene transfer of pigment epithelium-derived factor. Mol. Ther. 2003; 8: Involvement of BAI2 in the ischemia-induced brain angiogenesis. 72–9. J. Cereb. Blood Flow Metab. 2002; 22: 1054–67. 7 Matsumoto K, Ishikawa H, Nishimura D, Hamasaki K, Nakao K, 27 Soker S, Takashima S, Miao HQ, Neufeld G, Klagsbrun M. Eguchi K. Antiangiogenic property of pigment epithelium-derived factor Neuropilin-1 is expressed by endothelial and tumor cells as an in hepatocellular carcinoma. Hepatology 2004; 40: 252–9. isoform-specific receptor for vascular endothelial growth factor. Cell 8 Uehara H, Miyamoto M, Kato K, Ebihara Y, Kaneko H, Hashimoto H 1998; 92: 735–45. et al. Expression of pigment epithelium-derived factor decreases liver 28 Stephenson JM, Banerjee S, Saxena NK, Cherian R, Banerjee SK. metastasis and correlates with favorable prognosis for patients with Neuropilin-1 is differentially expressed in myoepithelial cells and ductal pancreatic adenocarcinoma. Cancer Res. 2004; 64: 3533–7. vascular smooth muscle cells in preneoplastic and neoplastic human 9 Guan M, Pang CP, Yam HF, Cheung KF, Liu WW, Lu Y. Inhibition of breast: A possible marker for the progression of breast cancer. Int. J. glioma invasion by overexpression of pigment epithelium-derived factor. Cancer 2002; 101: 409–14. Cancer Gene Ther. 2004; 11: 325–32. 29 Straume O, Akslen LA. Increased expression of VEGF-receptors 10 Halin S, Wikstrom P, Rudolfsson SH et al. Decreased pigment (FLT-1, KDR, NRP-1) and thrombospondin-1 is associated with epithelium-derived factor is associated with metastatic phenotype in glomeruloid microvascular proliferation, an aggressive angiogenic human and rat prostate tumors. Cancer Res. 2004; 64: 5664–71. phenotype, in malignant melanoma. Angiogenesis 2003; 6: 295– 11 Doll JA, Stellmach VM, Bouck NP et al. Pigment epithelium-derived 301. factor regulates the vasculature and mass of the prostate and pancreas. 30 Parikh AA, Fan F, Liu WB et al. Neuropilin-1 in human colon cancer: Nat. Med. 2003; 9: 774–80. Expression, regulation, and role in induction of angiogenesis. Am. J. 12 Isaacs JT. The biology of hormone refractory prostate cancer. Why does Pathol. 2004; 164: 2139–51. it develop? Urol. Clin. North Am. 1999; 26: 263–73. 31 Miao HQ, Lee P, Lin H, Soker S, Klagsbrun M. Neuropilin-1 13 Oh WK, Kantoff PW. Management of hormone refractory prostate expression by tumor cells promotes tumor angiogenesis and progression. cancer: Current standards and future prospects. J. Urol. 1998; 160: FASEB J. 2000; 14: 2532–9. 1220–9. 32 Blancher C, Harris AL. The molecular basis of the hypoxia response 14 Ishibashi K, Imai M. Prospect of a stanniocalcin endocrine/paracrine pathway: Tumour hypoxia as a therapy target. Cancer Metastasis Rev. system in mammals. Am. J. Physiol. Renal Physiol. 2002; 282: F367–7. 1998; 17: 187–94. 15 Yamamura J, Miyoshi Y, Tamaki Y et al. mRNA expression level of 33 Acker T, Diez-Juan A, Aragones J et al. Genetic evidence for a tumor estrogen-inducible gene, alpha 1-antichymotrypsin, is a predictor of suppressor role of HIF-2alpha. Cancer Cell. 2005; 8: 131–41. early tumor recurrence in patients with invasive breast cancers. Cancer 34 Guan M, Jiang H, Xu C, Xu R, Chen Z, Lu Y. Adenovirus-mediated Sci. 2004; 95: 887–92. PEDF expression inhibits prostate cancer cell growth and results in 16 Gagliardi AD, Kuo EY, Raulic S, Wagner GF, DiMattia GE. Human augmented expression of PAI-2. Cancer Biol. Ther. 2007; 6: 419–25. stanniocalcin-2 exhibits potent growth-suppressive properties in 35 Zhang T, Guan M, Xu C, Chen Y, Lu Y. Pigment epithelium-derived transgenic mice independently of growth hormone and IGFs. Am. J. factor inhibits glioma cell growth in vitro and in vivo. Life Sci. 2007; Physiol. Endocrinol. Metab. 2005; 288: E92–105. 81: 1256–63.