Science is the great antidote to the poison of enthusiasm ...shodhganga.inflibnet.ac.in/bitstream/10603/2604/14/14_appendix.pdfConference, Lonavala (MH) January 8-11th 2007. (Awarded

Science is the great antidote to the poison of enthusiasm and superstition

Adam Smith

108

9. APPENDIX 9.1 List of Publications 1. Hypoxia Regulates Cross-talk between Syk and Lck Leading to Breast Cancer Progression

and Angiogenesis. Goutam Chakraborty, Hema Rangaswami, Shalini Jain, and Gopal C.

Kundu. Journal of Biological Chemistry 2006: 281 (16) pp. 11322–11331.

2. The crucial role of cyclooxygenase-2 in osteopontin-induced protein kinase C α/c-Src/IκB

kinase α/β–dependent prostate rumor progression and angiogenesis. Shalini Jain, Goutam

Chakraborty, and Gopal C. Kundu. Cancer Research 2006; 66: (13) pp. 6638-6648.

3. The multifaceted roles of osteopontin in cell signaling, tumor progression and

angiogenesis. Shalini Jain, Goutam Chakraborty, Reeti Behera, Mansoor Ahmed, Priyanka

Sharma, Vinit Kumar and Gopal C. Kundu. Current Molecular Medicine 2006, 6 (8) pp.

819-830.

4. Osteopontin: an emerging therapeutic target for anticancer therapy. Shalini Jain, Goutam

Chakraborty, Anuradha Bulbule, Rajinder Kaur and Gopal C. Kundu. Expert Opinion on

Therapeutic Targets. 2007; 11(1) pp. 81-90.

5. Cyclooxygenase-2 (COX-2) plays crucial role in osteopontin-induced PKCα/IKKα/β

dependent prostate tumor growth and angiogenesis. Shalini Jain, Goutam Chakraborty, and

Gopal C. Kundu. Medimond International Proceeding 2007 pp. 105-112.

6. Hypoxia/Reoxygenation controls the crosstalk between Syk and Lck leading to regulation

of MelCAM-expression: signaling mechanism underlying breast cancer progression and

angiogenesis. Goutam Chakraborty, Shalini Jain and Gopal C. Kundu. Medimond

International Proceeding 2007 pp. 97-103.

7. Osteopontin-regulated signaling network in cancer: redefining the molecular targets.

Goutam Chakraborty, Shalini Jain, Reeti Behera, Mansoor Ahmed, Priyanka Sharma,

109

Vinit Kumar, Rajinder Kaur, Remya Raja, Smita Kale, Anuradha Bulbule and Gopal C.

Kundu. International Journal of Molecular Medicine 2007; 20 (S1) pp. S75.

8. Osteopontin promotes VEGF dependent breast tumor growth and angiogenesis via

autocrine and paracrine mechanisms. Goutam Chakraborty, Shalini Jain and Gopal C.

Kundu. Cancer Research 2008; 68 (1) pp. 152-161.

9. Down-regulation of osteopontin attenuates breast tumor progression in vivo. Goutam

Chakraborty, Shalini Jain, Tushar V. Patil, Goapl C. kundu. Journal of Cellular and

Molecular Medicine 2008; 12 (In press).

10. Curcumin abrogates osteopontin (OPN) induced VEGF dependent tumor angiogenesis.

Goutam Chakraborty, Shalini Jain, Gopal C. Kundu Molecular Medicine Report 2008;

(In press).

11. Prostaglandin E2 regulates tumor angiogenesis in prostate cancer. Shalini Jain, Goutam

Chakraborty, Ramya Raja, Smita Kale and Gopal C. Kundu (Manuscript under revision in

Cancer Research 2008; 68: 7750-7759.

12. Osteopontin regulates prostate tumor growth, metastasis and angiogenesis by modulating

the balance between NF-κB and FOXO3a. Shalini Jain, Goutam Chakraborty, Ramya

Raja, Smita Kale and Gopal C. Kundu. (Manuscript under preparation).

110

9.2 Conference Presentations 1. Oral presentation entitled “osteopontin regulates prostate tumor growth, metastasis and

angiogenesis by modulating the balance between NF-κB and FOXO3a” was given by

Shalini Jain, Goutam Chakraborty and Gopal C. Kundu in “2nd International

Symposium on Translational Research, Natural products and cancer” December 9-

12, 2007 Lonavala India.

2. Oral presentation entitled “ (COX-2) and its metabolites results in drastic reduction of

OPN-induced tumor progression: a new dimension in the therapeutics of prostate

cancer” was give by Shalini Jain, Goutam Chakraborty, Gopal kundu in 26th Annual

Convention of Indian Association for Cancer Research and International

Symposium on Translational Research in Cancer, January 17th-19th 2007,

Bhuvaneshwar, India.

3. Poster entitled “cyclooxygenase-2 (COX-2) plays crucial role in osteopontin-induced

PKCα/IKKα/β dependent prostate tumor growth and angiogenesis” by Shalini Jain,

Goutam Chakraborty, and Gopal C. Kundu was presented in the 3rd SFRR-Asia

Conference, Lonavala (MH) January 8-11th 2007. (Awarded as the Best Poster)

4. Poster entitled “Hypoxia/Reoxygenataion controls the crosstalk between Syk and Lck

leading to regulation of MelCAM-expresion: signaling mechanism underlying breast

cancer progression and angiogenesis” by Goutam Chakraborty, Shalini Jain, Gopal C.

Kundu was presented in 3rd Biennial Meeting of the Society for Free Radical

Research-Asia-SFRR-Asia, January 8-11, 2007, Lonavala, India

5. Poster entitled “suppression of cyclooxygenase-2 (COX-2) and its metabolites results in

drastic reduction of osteopontin-induced prostate cancer progression” by Shalini Jain,

Goutam Chakraborty and Goapl C. Kundu was presented in XXIX All India Cell

Biology Conference & Symposium onGene to Genome: Environment and

Chemical Interaction, Lucknow (U.P.), January 17-20, 2006.

The Crucial Role of Cyclooxygenase-2 in Osteopontin-Induced

Protein Kinase C A/c-Src/IKB Kinase A/B–Dependent Prostate

Tumor Progression and Angiogenesis

Shalini Jain, Goutam Chakraborty, and Gopal C. Kundu

National Center for Cell Science, Pune, India

Abstract

The regulation of tumor progression towards its malignancyneeds the interplay among several cytokines, growth factors,and enzymes, which are controlled in the tumor microenvi-ronment. Here, we report that osteopontin, a small integrin-binding ligand N-linked glycoprotein family of calcifiedextracellular matrix–associated protein, regulates prostatetumor growth by regulating the expression of cyclooxygenase-2 (COX-2). We have shown that osteopontin stimulates theactivation of protein kinase C A/nuclear factor–inducingkinase/nuclear factor-KB–dependent signaling cascades thatinduces COX-2 expression, which in turn regulates the pro-staglandin E2 production, matrix metalloproteinase-2 activa-tion, and tumor progression and angiogenesis. We haverevealed that suppression of osteopontin-induced COX-2 ex-pression by the nonsteroidal anti-inflammatory drug celecoxibor blocking the EP2 receptor by its blocking antibody resultedin significant inhibition of cell motility and tumor growth andangiogenesis. The data also showed that osteopontin-inducedmice PC-3 xenograft exhibits higher tumor load, increasedtumor cell infiltration, nuclear polymorphism, and neovascu-larization. Interestingly, use of celecoxib or anti-EP2 blockingantibody drastically suppressed osteopontin-induced tumorgrowth that further indicated that suppression of COX-2 orits metabolites could significantly inhibit osteopontin-inducedtumor growth. Human clinical prostate cancer specimenanalysis also supports our in vitro and animal model studies.Our findings suggest that blockage of osteopontin and/orCOX-2 is a promising therapeutic approach for the inhibitionof prostate tumor progression and angiogenesis. (Cancer Res2006; 66(13): 6638-48)

Introduction

Cancer of the prostate is the most frequently diagnosed cancerand one of the major causes of death in men especially in westernworld (1). Prostate cancer progression is a series of complex events,which require crosstalk between several oncogenic molecules, andenable the cancer to spread and evoke angiogenesis. A high level ofconstitutive cyclooxygenase-2 (COX-2) expression has beendetected in colorectal, gastric, pancreatic, head and neck, lung,breast, and in other cancers (2). COX is an integral membranebifunctional enzyme, which metabolizes arachidonic acids to manybiologically active eicosanoids (3–5). There are three isoforms of

the COX enzyme (COX-1, COX-2, and COX-3). COX-1 is involved inthe production of prostaglandin (PG)–mediating cellular andphysiologic functions. The functional role of COX-3 in humanphysiology and pathophysiology remains to be established (6).

COX-2 is induced in many cell types by mitogens, growth factors,cytokines, and tumor promoters, and its increased expression isassociated with cancer progression through a PG-dependentmanner (2–5). COXs convert free arachidonic acids into PGs andthromboxanes (3). In human prostate carcinoma, significant levelsof PGE2 has been detected, which plays an important role in cancerprogression (7). PGE2 can contribute to tumor developmentthrough several mechanisms, including promotion of angiogenesis,inhibition of apoptosis, increased invasiveness and motility, andmodulation of inflammation and immune responses (8–10). Recentreports showed that COX-2 promotes tumor cell proliferation,survival, and angiogenesis through a PGE2-mediated pathway(10, 11).

Osteopontin, a secreted, noncollagenous, chemokine-like smallintegrin-binding ligand N-linked glycoprotein family of proteinplays significant role in determining the oncogenic potential ofvarious cancers and recognized as a key marker in the processes oftumorigenicity and metastasis (12). Osteopontin is an acidicglycoprotein with a molecular mass varying from 44 to 75 kDa,depending on the degree of posttranslational modification (13). Ithas an NH2-terminal signal sequence, a sialic acid region consistingof nine consecutive aspartic acid residues, a GRGDS cell adhesionsequence that is predicted to be flanked by h-sheet structure (14).Osteopontin interacts with several integrins and CD44 variants inan RGD sequence–dependent and RGD sequence–independentmanner (15, 16). Osteopontin is involved in normal tissueremodeling processes, such as bone resorption, wound healing,and tissue injuries as well as restenosis, atherosclerosis, tumor-igenesis, and autoimmune diseases (17, 18). Previous studiesindicated that osteopontin plays crucial role in chemotaxis andchemoinvasion of PC-3 cells (19). Recent data also showed thatosteopontin induces pro-matrix metalloproteinase-2 (pro-MMP-2)and pro-MMP-9 activation, urokinase plasminogen activator (uPA)secretion, cell motility, extracellular matrix invasion, and tumorgrowth (20, 21).

Cell motility, a major step in cancer metastasis, is oftenassociated with the activation of protein tyrosine or serine kinases,like c-Src and protein kinase C a (PKCa), respectively. Previousstudies have indicated that c-Src and PKCa play crucial rolesin COX-2 expression and COX-2-dependent tumor progression(22–24). The transcription factor nuclear factor-nB (NF-nB) act as akey molecule in regulating wide range of physiologic andpathologic processes (25). Nuclear factor inducing kinase (NIK), amember of mitogen-activated protein kinase kinase kinase(MAPKKK) family has been reported to activate NF-nB throughphosphorylation and degradation of InBa (26). We have recently

Requests for reprints: Gopal C. Kundu, National Center for Cell Science, Pune 411007, India. Phone: 91-20-25690922 ext. 1103; Fax: 91-20-25692259; E-mail: [email protected].

I2006 American Association for Cancer Research.doi:10.1158/0008-5472.CAN-06-0661

Cancer Res 2006; 66: (13). July 1, 2006 6638 www.aacrjournals.org

Research Article

reported that osteopontin induces uPA secretion and MMP-2/MMP-9 activation through c-Src/phosphatidylinositol 3-kinase/MAPK signaling pathways (20, 21, 27). However, the molecularmechanism by which osteopontin regulates PKCa/c-Src–depen-dent InB kinase (IKK)–mediated NF-nB activation, which ulti-mately regulates tumor progression and angiogenesis throughinduction of COX-2 expression in prostate cancer and signalingcascades, underlying these processes are not well defined.

In this study, we provide both in vitro and in vivo experimentalevidences, at least in part, the molecular mechanism by whichosteopontin regulates PKCa/c-Src/IKK/NF-nB signaling cascadesleading to COX-2-mediated PGE2 production and MMP-2 activa-tion in prostate cancer. Furthermore, we have shown thatosteopontin-induced COX-2 regulates cell motility, angiogenesis,and tumorigenesis of prostate cancer through both autocrine andparacrine pathways. However, suppression of COX-2 activity bynonsteroidal anti-inflammatory drug (NSAID) celecoxib or blockingthe interaction between PGE2 and its receptor EP2 by using specificanti-EP2 blocking antibody significantly suppressed osteopontin-induced in vitro cell motility, invasiveness, and in vivo tumorgrowth. Moreover, the clinical data indicated that the increasedexpressions of osteopontin and COX-2 correlate with enhancedMMP-2 expression and angiogenesis in prostate cancer specimensof higher grades. Consequently, osteopontin plays important andessential role in two key aspects of tumor progression: COX-2-mediated PGE2 production and MMP-2 activation by tumor cellsleading to tumor progression and COX-2/PGE2–stimulated angio-genesis. Our findings suggest that blockade of osteopontin andCOX-2 is a promising therapeutic approach for the inhibition oftumor progression by suppressing tumor growth and angiogenesis.

Materials and Methods

Antibodies and reagents. Rabbit polyclonal anti-COX-2, anti-PKCa,anti-NIK, anti-p-NIK(Thr559), anti-IKKa/h, anti-NF-nB, p65, anti-MMP-2,

anti-c-Src, anti-EP2, and anti-actin, mouse monoclonal anti-phosphotyr-

osine, goat polyclonal anti-phospho-PKCa (Ser657) antibodies were

purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbitpolyclonal anti-p-IKKa/h (Ser180/Ser181) antibody was from Cell Signaling

Technology (Beverly, MA); goat polyclonal anti-osteopontin antibody was

from R&D Systems (Minneapolis, MN); and rabbit polyclonal anti-humanvWF antibody was purchased from Sigma (St. Louis, MO). The [g-32P]ATPwas purchased from Board of Radiation and Isotope Technology

(Hyderabad, India). All other chemicals were of analytic grade. The human

osteopontin was purified from milk as described previously (20), with aminor modification, and used throughout this study.

Cell culture. The human prostate cancer (PC-3) cells were obtained

from the American Type Culture Collection (Manassas, VA). Human

umbilical vein endothelial cell line EA.hy-926 was a generous gift from Dr.Christopher Newton (University of Hull, United Kingdom). PC-3 cells were

cultured in F-12 HAM Nutrient Mixture (Sigma), and EA.hy-926 cells were

cultured in DMEM (Sigma) supplemented with 10% FCS, 100 units/mL

penicillin, 100 Ag/mL streptomycin, and 2 mmol/L glutamine in ahumidified atmosphere of 5% CO2 and 95% air at 37jC.

Plasmids and DNA transfection. The full-length osteopontin cDNA was

a generous gift from Dr. Ann Chambers (University of Western Ontario,London, Ontario, Canada). The dominant-negative c-Src (pcSrcY527F/S12A)

and COX-2-Luc (�1432/+59) constructs were generous gifts from Dr. David

Shalloway (Cornell University, NY) and Dr. Hiroyashu Inoue (Osaka, Japan;

ref. 28), respectively. The wild-type NIK (wt pcDNA NIK) and kinase-negative NIK (mut pc DNA NIK and NIK-K429A/K430A) in pcDNA3 were

obtained from Prof. David Wallach (Weizmann Institute of Science,

Rehovot, Israel). The wild-type and dominant-negative constructs of IKKa(wt IKKa and dn IKKa) and IKKh (wt IKKh and dn IKKh) in

PKC-related kinase were kind gifts from Prof. D.V. Goeddel (Tularik, Inc.,San Francisco, CA). The super-repressor form of InBa cDNA fused

downstream to a FLAG epitope in an expression vector (pCMV4) was a

gift from Dr. Dean Ballard (Vanderbilt University School of Medicine,

Nashville, TN). PC-3 cells were split 16 hours before transfection andtransiently transfected with cDNA using LipofectAMINE 2000 reagent

according to manufacturer’s instructions (Invitrogen, San Diego, CA). The

cell viability was detected by trypan blue dye exclusion test. Transfected

cells were used for COX-2 expression, COX-2 promoter analysis studies, IKKand p65 phosphorylation, and PKCa kinase assay and migration and

invasion assays.

Small interfering RNA. PC-3 cells were transfected with small

interfering RNA (siRNA) that specifically targets the osteopontin gene ornonsilencing control using LipofectAMINE-2000 according to manufac-

turer’s instructions. siRNA duplexes were synthesized by Dharmacon, Inc.

(Lafayette, CO). The sequence targeted for osteopontin is 5¶-GUUUCA-CAGCCACAAGGACdTdT/dTdTCAAAGUGUCGGUGUUCCUG-5¶, and the

nonsilencing control is 5¶-CAGUACAACGCAUCUGGCAdTdT/dTdTGUCAU-GUUGCGUAGACCGU-5¶.

Western blot and immunoprecipitation. The level of COX-2, PKCac-Src, osteopontin, IKKa/h, actin expressions and phosphorylations of

IKKa/h and NF-nB, p65 in transfected or treated PC-3 cells were analyzed

by Western blot using its specific antibody (20). The tyrosine phosphory-

lation of IKKa/h and the interaction between PKCa and c-Src were done byimmunoprecipitation followed by Western blot.

RNA extraction and reverse transcription-PCR. Total RNA was

isolated from osteopontin-treated PC-3 cells and analyzed by reversetranscription-PCR (RT-PCR). The reverse transcription and PCR amplifica-

tion used 10 Ag of total RNA, with Moloney murine leukemia virus reverse

transcriptase (Invitrogen) and primers COX-2 forward (5¶-TTCAAATGA-GATTGTGGGAAAATTGCT-3¶) and COX-2 reverse (5¶-AGATCATCT-CTGCCTGAGTATCTT-3¶) to reverse transcribe sense and antisense RNAs,

respectively (29). The amplified cDNA fragments were resolved by 2%

agarose gel electrophoresis.

COX-2 luciferase assay. The semiconfluent PC-3 cells were grown in 24-well plates and transiently transfected with COX-2 luciferase reporter

construct using LipofectAMINE 2000. In separate experiments, cells were

individually transfected with wt and mut NIK and wt and dn IKKh alongwith COX 2 luciferase construct. The transfection efficiency was normalized

by cotransfecting the cells with pRL vector (Promega, Madison, WI)

containing a full-length Renilla luciferase gene under the control of a

constitutive promoter. After 24 hours of transfection, the cells were treatedwith 0.5 Amol/L osteopontin, and the luciferase activities were measured by

luminometer (Thermo Electron) using the dual luciferase assay system

according to the manufacturer’s instructions (Promega). Changes in the

luciferase activity with respect to control were calculated.Estimation of PGE2. The levels of PGE2 from the conditioned media of

osteopontin-treated PC-3 cells and mice plasma were detected by using the

Biotrack PGE2 competitive enzyme immunoassay kit (Amersham Bioscien-

ces, Piscataway, NJ) according to the manufacturer’s instructions.In vitro kinase assay by autophosphorylation study. In vitro kinase

assay was done as described earlier (30). Briefly, PC-3 cells were either

treated with osteopontin (0.5 Amol/L) for 5 minutes or pretreated with 100

nmol/L staurosporine (Calbiochem, La Jolla, CA) for 3 hours or transfected

with dn c-Src and then treated with osteopontin. Cell lysates were

immunoprecipitated with anti-PKCa and anti-c-Src antibodies. Autophos-

phorylation of PKCa and c-Src were measured by incubating the

immunoprecipitated samples with 40 ACi of [g32P]ATP in 40 AL kinase

assay buffer [20 mmol/L HEPES (pH 7.7), 2 mmol/L MgCl2, 10 mmol/L

h-glycerophosphate, 10 mmol/L NaF, 10 mmol/L pNPP, 300 Amol/L Na3VO4,

1 mmol/L benzamidine, 2 Amol/L phenylmethylsulfonyl fluoride, 10 Ag/mL

aprotinin, 1 Ag/mL leupeptin, 1 Ag/mL pepstatin, and 1 mmol/L DTT] for

30 minutes at 30jC. Reactions were terminated by the addition of sample

buffer and analyzed by SDS-PAGE followed by autoradiography.Zymography. The gelatinolytic activity was measured as described

previously (20). Briefly, PC-3 cells were either treated with osteopontin

or pretreated with staurosporine and COX-2 inhibitor (Celecoxib) or

COX-2 and Tumor Progression and Angiogenesis

www.aacrjournals.org 6639 Cancer Res 2006; 66: (13). July 1, 2006

transfected with dn c-Src and then treated with osteopontin. Conditionedmedium was collected, and gelatinolytic activity of MMP-2 was detected by

zymography.

Cell migration and invasion assay. The migration and invasion assays

were done using Transwell cell culture chamber (Corning, Corning, NY) andMatrigel coated invasion chambers (Collaborative Biomedical, Bedford,

MA), respectively, according to the standard procedure as described

previously (27). Briefly, the confluent monolayers of PC-3 (treated or

transfected) cells were harvested with trypsin-EDTA and centrifuged at 800� g for 10 minutes. The cell suspension (2 � 105 per well) was added to the

upper chamber. The lower chamber was filled with fibroblast-conditioned

medium, which acted as a chemoattractant. In separate experiments,

osteopontin-treated conditioned medium of PC-3 cells was incubated withEA.hy-926 cells alone or along with anti-EP-2 blocking antibody and

subjected to migration and invasion assays. After 12 hours, cells in the lower

chamber were fixed and stained with Giemsa stain and counted in threehigh-power fields (C/HPF) under an inverted microscope (Olympus). Data

are presented as the average of three counts FSE.

Wound assay. The wound assay was done as described (31). Briefly, the

post-confluent PC-3 cells with typical cobblestone morphology were used inthis experiment. Wounds with a constant diameter were made. Cells were

treated with osteopontin alone or along with EP2 blocking antibody or

celecoxib (COX-2 inhibitor). In other experiments, EA.hy-926 cells were

incubated with conditioned media obtained from osteopontin-treated PC-3cells either alone or along with EP2 blocking antibody, and wound assay was

conducted. The wound photographs were taken under phase-contrast

microscope (Olympus).PC-3 xenograft tumor model. The tumorigenicity experiments were

done as described previously (27). Briefly, PC-3 cells (5 � 105) were mixed

with an equal volume of cold Matrigel and then injected s.c. into the dorsal

side of the athymic nude mice (NMRI, nu/nu ; NIV, Pune, India). Osteopontinalone or along with anti-EP-2 blocking antibody (20 Ag) was injected into

tumor sites twice a week for up to 4 weeks. In other experiments, celecoxib

(1,500 ppm) was given along with diet of the osteopontin-injected animalsas described earlier (32). Three mice were used in each set of experiments.

The mice were kept under pathogen-free conditions. Growths of s.c. tumors

were monitored weekly by measuring the tumors with calipers. At the

termination of the experiment, blood was collected from the retro-orbitalplexus under anesthesia from both experimental and control groups.

Animals were sacrificed by cervical dislocation, and tumors were excised

and weighed. The tumor samples were used for histopathologic and

immunohistochemical studies using standard procedure (33). The expres-sions of COX-2, MMP-2, and vWF (endothelial cell–specific marker) were

determined by immunofluorescence using their specific antibodies. The

slides were analyzed under confocal microscopy (Zeiss, Jena, Germany).

Human prostate clinical sample analysis. Human prostate tumorspecimens of different Gleason grades and normal prostate tissues were

collected from a local hospital with informed consent. The COX-2,

osteopontin, MMP-2, and vWF expressions were detected by immunoflu-orescence using specific antibodies. Five samples of each group [normal,

low grade (prostatic intraepithelial neoplasia or PIN), and malignant] were

analyzed. Normal prostate tissues were used as control.

Statistical analysis. The data reported in cell migration, wound healing,and invasion are expressed as mean FSE. Statistical differences were

determined by two-way ANOVA and Student’s t test. P < 0.05 was considered

significant. All these bands were analyzed densitometrically (Kodak Digital

Science, Rochester, NY), and the fold changes were calculated.

Results

Osteopontin augments COX-2 protein and mRNA expres-sion. To determine whether osteopontin augments COX-2expression, PC-3 cells were treated with 0.5 Amol/L osteopontinfor 0 to 24 hours. Expression of COX-2 in cell lysates wasdetected by Western blot, and the data indicated that osteopontin

Figure 1. Osteopontin (OPN ) induces COX-2 promoter activity and expression in PC-3 cells. A, serum-starved PC-3 cells were treated with 0.5 Amol/L osteopontin for0 to 24 hours. The level of COX-2 from whole-cell lysates was analyzed by Western blot. Actin was used as loading control. B, cells were treated with osteopontin(0-0.5 Amol/L) for 6 hours; total RNA was isolated; and RT-PCR analysis was done using COX-2-specific primers. Changes in mRNA levels were determined asfold of induction. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as internal control. C, PC-3 cells were transfected with osteopontin cDNA,osteopontin-specific siRNA, or nonsilencing scrambled siRNA. In separate experiments, osteopontin siRNA–transfected cells were either cotransfected with osteopontincDNA or treated with 0.5 Amol/L osteopontin. COX-2 was analyzed by Western blot. The level of osteopontin is also detected by Western blot. Actin was used as loadingcontrol. D, osteopontin enhances NIK/IKK–mediated COX-2 promoter activity. Cells were transfected with luciferase reporter construct (COX-2-Luc) and thentreated with 0 to 0.5 Amol/L osteopontin or cotransfected with wt and mut NIK or wt and dn IKKh and then treated with 0.5 Amol/L osteopontin. Cell lysates were used tomeasure the luciferase activity. The values were normalized to Renilla luciferase activity. The fold changes were calculated. Columns, mean of triplicate determinations;bars, SE.

Cancer Research


induced maximum COX-2 expression at 24 hours (Fig. 1A). Thedose-dependent effect of osteopontin (0-1 Amol/L) on COX-2expression was also detected, and the results indicated thatosteopontin at 0.5 Amol/L exhibits maximum level of COX-2expression in PC-3 cells (data not shown). The COX-2 mRNA levelwas also determined by RT-PCR, and the data indicated thatosteopontin at 0.5 Amol/L enhanced maximum COX-2 expressionat transcriptional level (Fig. 1B). Interestingly, our results alsoshowed that osteopontin induced COX-2 expression in otherprostate cancer cell lines like DU-145 and LNCaP (data not shown).To examine the specificity of osteopontin on COX-2 expression,cells were transfected with wt osteopontin cDNA alone orcotransfected with osteopontin siRNA duplex and control RNAduplex. The data showed that cells transfected with osteopontincDNA enhanced the COX-2 expression. Cells transfected withosteopontin cDNA followed by siRNA duplex suppressed theosteopontin-induced COX-2 expression (Fig. 1C). Our resultsindicated that osteopontin enhanced COX-2 expression both attranscriptional and translational levels.Osteopontin enhances NIK/IKK/NF-KB–dependent COX-2

promoter activity. The effect of osteopontin on COX-2 promoteractivity was determined by transfecting PC-3 cells with COX-2luciferase construct followed by treatment with osteopontin in adose-dependent manner. In separate experiments, cells werecotransfected with wt NIK or IKKh or kinase-negative (mut) NIKor dn IKKh along with COX-2 luciferase construct and then treated

with osteopontin. Further transfecting the cells with pRL constructnormalized the transfection efficiency, and COX-2 luciferaseactivities were measured according to standard procedure. Theresults showed that cells transfected with wt NIK or IKKhenhanced the COX-2 promoter activity, whereas mut NIK or dnIKKh significantly suppressed osteopontin-induced COX-2 pro-moter activity (Fig. 1D). Cells transfected with wt IKKa enhancedosteopontin-induced COX-2 promoter activity, whereas dn IKKa orthe super-repressor form of InBa suppressed osteopontin-inducedCOX-2 promoter activity (data not shown). These results indicatedthat osteopontin enhanced COX-2 promoter activity through aNIK/IKK–dependent NF-nB-mediated pathway.Osteopontin regulates PKCA/c-Src–dependent IKK/NF-KB–

mediated COX-2 expression. Recent reports revealed that PKCaplays important role in the survival and growth of androgen-independent human prostate cancer (PC-3) cells (34). It is wellestablished that PKCa plays crucial role in COX-2 expression (23).Therefore, to delineate the effect of osteopontin on PKCaphosphorylation and its interaction with c-Src, cells were treatedwith osteopontin for 0 to 7 minutes, and one part of the cell lysateswas analyzed by Western blot using anti-phospho-PKCa antibody(Fig. 2A, a, top). The remaining part of lysates was immunopre-cipitated with anti-c-Src antibody, and the interaction betweenphosphorylated PKCa and c-Src was determined by Western blotusing anti-phospho-PKCa antibody (Fig. 2A, b, top). The same blotswere reprobed with anti-PKCa and c-Src antibody, respectively, as

Figure 2. Osteopontin (OPN ) induces PKCa activation, PKCa-mediated c-Src and IKK phosphorylation, and COX-2 expression. A, PC-3 cells were treated withosteopontin for 0 to 7 minutes, and the phosphorylation of PKCa was detected by Western blot using anti-phospho-PKCa antibody (p-PKCa ; a, top ). The cellswere treated with osteopontin for 0 and 5 minutes, and cell lysates were immunoprecipitated with anti-c-Src antibody and analyzed by Western blot usinganti-phospho-PKCa antibody (b, top ). Levels of PKCa and c-Src were detected by reprobing the same blots with their specific antibodies (a and b, bottom ). B, cellswere either transfected with dn c-Src or pretreated with staurosporine (Stau , a PKC inhibitor, 100 nmol/L) and then treated with osteopontin for 5 minutes. Celllyates were immunoprecipitated with anti-PKCa or anti-c-Src antibody, and the kinase activity of PKCa (a) and c-Src (b ) was determined by autophosphorylationexperiment and visualized by autoradiography. C, cells were individually transfected with dn c-Src, wt and mut NIK, or pretreated with staurosporine and then treatedwith osteopontin. Half of the cell lysates were analyzed by Western blot using anti-phospho-Ser-IKKa/h (Ser180/Ser181) antibody (top), and the remaining part of celllysates was immunoprecipitated with IKKa/h antibody and detected by Western blot using anti-phosphotyrosine antibody (middle ). The level of IKKa/h was alsodetected by Western blot (bottom ). D, cells were pretreated with staurosporine or individually transfected with dn c-Src, InBa super-repressor (sup. rep. ), wt and dnIKKa or IKKh, and wt or mut NIK and then treated with osteopontin. COX-2 expression was detected by Western blot. Actin was used as loading control. Representsthree experiments exhibiting similar results. Fold changes were calculated.



loading controls (Fig. 2A, a and b, bottom). The results indicatedthat osteopontin enhanced PKCa phosphorylation and its interac-tion with c-Src. To determine the role of avh3 integrin onosteopontin-induced PKCa phosphorylation, cells were pretreatedwith anti-human avh3 integrin blocking antibody, RGD, or RGE(GpenGRGDSPCA and GRGESP, respectively) peptides and thentreated with osteopontin, and the level of PKCa phosphorylation wasdetected by immunoblotting. The results indicated that anti-avh3

integrin blocking antibody and GpenGRGDSPCA but not GRGESPsuppressed the osteopontin-induced PKCa phosphorylation (datanot shown). The data suggested that avh3 integrin plays animportant role in osteopontin-induced PKCa phosphorylation. Todelineate whether PKCa is upstream of c-Src or vice versa, cells werepretreated with PKCa inhibitor (staurosporine) and treated withosteopontin, and cell lysates were immunoprecipitated with anti-c-Src antibody. The immunoprecipitated samples were analyzed forc-Src autophosphorylation by in vitro kinase assay. Similarly, in otherexperiments, cells were transfected with dn c-Src and treated withosteopontin, and cell lysates were immunoprecipitated with anti-PKCa. The immunoprecipitated samples were used for in vitro PKCakinase assay. The data showed that dn c-Src has no effect on PKCaautophosphorylation, whereas PKCa inhibitor suppressed c-Srcphosphorylation, suggesting that osteopontin regulates PKCa-dependent c-Src activation (Fig. 2B, a and b, top). The levels ofPKCa and c-Src in the same lysates were detected by Western blotusing their specific antibodies as controls (Fig. 2B, a and b, bottom).Furthermore, suppression of PKCa and c-Src resulted in inhibition ofosteopontin-induced NIK phosphorylation (data not shown). Todetermine whether these upstream kinases (PKC, c-Src, and NIK)play any role in osteopontin-induced IKKa/h phosphorylation, cellswere individually transfected with dn c-Src or wt and mut NIK orpretreated with PKCa inhibitor (staurosporine) and then treatedwith osteopontin. Cell lysates were analyzed either by Western blot

using anti-phosphoserine IKKa/h antibody or immunoprecipitatedwith anti-IKKa/h antibody followed by immunoblotting withanti-phosphotyrosine antibody (Fig. 2C, top and middle). Thesedata showed that both PKC inhibitor and dn c-Src suppressedosteopontin-induced serine and tyrosine phosphorylations of IKKa/h, whereas mut NIK inhibits only the serine but not tyrosinephosprylation of IKKa/h, suggesting that NIK plays differential rolein osteopontin-induced PKCa/c-Src–mediated IKKa/h activation(Fig. 2C, top and middle). The level of IKK a/h was also detected byWestern blot as control (Fig. 2C, bottom).

To further investigate whether PKCa, c-Src, NIK, IKK, and NF-nBplay any role in osteopontin-induced COX-2 expression, cells werepretreated with PKC inhibitor (staurosporine) or individuallytransfected with dn c-Src, wt and mut NIK, and InBa super-repressor, wt and dn IKKa or IKKh followed by treatment withosteopontin. The COX-2 expression was detected by Western blot(Fig. 2D). These results suggested that PKCa and c-Src play animportant role in regulating osteopontin-induced COX-2 expres-sion through a NIK/IKK/NF-nB–mediated pathway.Osteopontin induces PKCA/c-Src–mediated p65 phosphor-

ylation. Recent findings suggested that phosphorylation of p65subunit of NF-nB leads to nuclear translocation and activation ofNF-nB (35), and that the promoter region of COX-2 contains theNF-nB response element (36). In Fig. 2D , we showed thatsuppression of NF-nB by overexpression of InBa super-repressorsignificantly inhibits osteopontin-induced COX-2 expression.Therefore, we sought to determine whether osteopontin regulatesNF-nB, p65 phosphorylation that ultimately regulates COX-2expression, and whether PKCa and c-Src are involved in thisprocess. Accordingly, cells were treated with osteopontin from 0 to180 minutes or pretreated with PKC inhibitor (staurosporine) ortransfected with dn c-Src and then treated with osteopontin. Thelevel of p65 phosphorylation in the cell lysates was detected by

Figure 3. Osteopontin (OPN ) inducesPKCa/c-Src–dependent NF-nB, p65phosphorylation, MMP-2 activation, andPGE2 production. A, PC-3 cells weretreated with osteopontin for the indicatedtime, and the level of phospho-p65 (p-p65 ;Ser536) from cell lysates were detected byWestern blot (top ). B, cells weretransfected with dn c-Src or pretreated withstaurosporine (Stau ) and then treatedwith osteopontin. The level of phospho-p65was analyzed by Western blot. The level ofnon-phospho-p65 was also detected byWestern blot (A and B, bottom ). C, cellswere transfected with dn c-Src orpretreated with staurosporine or celecoxib(COX-2 inhibitor, 50 Amol/L) and thentreated with osteopontin for 24 hours.Conditioned media were collected, and thelevel of MMP-2 activation was detected byzymography. Recombinant MMP-2(r MMP-2) was used as positive control.D and E, cells were treated with 0.5 Amol/Losteopontin for 0 to 24 hours. In separateexperiments, cells were transfected withdn c-Src or pretreated with staurosporineand celecoxib and then treated withosteopontin for 24 hours. Conditionedmedia were collected, and the level ofPGE2 was estimated by enzymeimmunoassay. Represents threeexperiments exhibiting similar results.

Cancer Research


Western blot using anti-phospho-p65 antibody (Fig. 3A and B). Theresults showed that inhibition of PKCa and c-Src significantlysuppressed osteopontin-induced NF-nB, p65 phosphorylation(Fig. 3B). These results clearly suggested that osteopontin inducesp65 phosphorylation through a PKCa/c-Src–dependent pathway,which leads to NF-nB activation that ultimately regulates COX-2expression.Osteopontin augments COX-2-mediated PGE2 production

and MMP-2 activation. Earlier reports revealed that overexpres-sion of COX-2 leads to up-regulation of arachidonic acidmetabolism, which in turn results in enhancement of PGE2production and tumor progression (2, 37–39). Accordingly, todetermine whether osteopontin can induce COX-2-mediated PGE2

production, cells were treated with osteopontin for 0 to 24 hours,and the PGE2 level in the conditioned media was measured byusing PGE2 enzyme immunoassay kit. To examine whether PKC,c-Src, and COX-2 are involved in osteopontin-induced PGE2

production, cells were pretreated with PKC inhibitor (staurospor-ine) or COX-2 inhibitor (celecoxib) or transfected with dn c-Src andthen treated with osteopontin. The results indicated that inhibitorsof PKC and COX-2 and dn c-Src drastically suppressed theosteopontin-induced PGE2 production (Fig. 3D and E). To delineatethe role of PKCa, c-Src, and COX-2 in osteopontin-induced MMP-2activation, cells were treated or transfected as described earlier,and the level of active MMP-2 in the conditioned medium wasdetermined by zymography (Fig. 3C). These results showed thatCOX-2 plays a crucial role in osteopontin-induced PKC/c-Src–mediated PGE2 production and MMP-2 activation in prostatecancer cells.COX-2 and PGE2 play an important role in osteopontin-

induced PC-3 cell migration and invasion. Previous datasuggested that COX-2 and its metabolite PGE2 act as crucialmolecules in regulating tumor cell motility, invasiveness, and

tumor metastasis (2). Therefore, to examine the role of COX-2 andPGE2 in osteopontin-mediated PC-3 cell migration, the woundassay was done. Cells were either pretreated with PKC inhibitor(staurosporine), celecoxib, or anti-EP2 antibody or transfected withdn c-Src. Wounds with a constant diameter were made, and cellswere treated with osteopontin. The wound photographs were takenunder phase-contrast microscope (Fig. 4A). These data showedthat inhibition of COX-2 or its upstream kinases (PKCa and c-Src)resulted in significant suppression of osteopontin-induced PC-3cell motility towards the wound and further indicated that COX-2act as a key molecule in osteopontin-induced tumor cell migration.Interestingly, blocking the interaction between PGE2 and itsreceptor EP2 by using specific anti-EP2 blocking antibody alsosuppressed osteopontin-induced tumor cell motility, indicatingthat osteopontin regulates wound migration via COX-2-dependentPGE2 production through interaction with its receptor EP2 in anautocrine fashion. The wound migration data were furtherconfirmed by modified Boyden chamber migration and extracel-lular matrix invasion assays. The results indicated that the cellmigration due to over expression of wt NIK, IKKa, or IKKh inresponse to osteopontin was suppressed by the celecoxib,suggesting the potential role of COX-2 in osteopontin-inducedNIK/IKK–dependent tumor cell migration and invasion (Fig. 4Band C). To examine the specificity of COX-2 in osteopontin-inducedcell migration, NS398 (NSAID), a COX-2 inhibitor, was used, and thedata showed that NS398 is also suppressed osteopontin-inducedcell migration and invasion. PGE2 was used as a positive control.Taken together, these data indicated the potential and crucial roleof COX-2 and PGE2 in osteopontin-induced tumor cell motility andinvasiveness.Osteopontin-induced tumor cell–derived PGE2 enhanced

endothelial cell motility and invasiveness through paracrinemechanism. Earlier data indicated that tumor-derived PGE2

Figure 4. Inhibition of the PKC, c-Src, andCOX-2 or blocking of EP2 receptorsuppressed osteopontin (OPN )–inducedPC-3 cell motility and invasiveness.A, cells were transfected with dn c-Src,or pretreated with staurosporine (Stau ),celecoxib, or anti-EP2 blocking antibody(4 Ag) then treated with osteopontin for12 hours. Wound assays were done asdescribed earlier. Wound photographswere taken at initial time (0 hour) and thetermination of the experiments (12 hours).The experiments were repeated intriplicate. B, cells were individuallytransfected with wt NIK and IKKa and IKKhand treated with celecoxib. In separateexperiments, cells were transfected withdn c-Src or treated with staurosporine,celecoxib, NS398, and EP2 blockingantibody and used for migration assay.Osteopontin was used in the upperchamber. PGE2 was used as positivecontrol. After 12 hours, migrated cells werecounted from the lower chamber andrepresented graphically. C, experimentsparalleling those of (B) but measuringinvasion rather than migration. Columns,means of three determinations; bars, SE.



interacts with the endothelial cell surface receptor and inducedendothelial cell motility and invasiveness that leads to tumorangiogenesis (38). It is reported that EP2 act as one of the mainPGE2 receptor in endothelial cells (39). To determine whetherosteopontin-induced PC-3 cell–derived PGE2 enhances endothelialcell migration and invasion through EP2-mediated paracrinemanner, endothelial cells were used on the upper side of modifiedBoyden or Matrigel-coated invasion chamber. The conditionedmedia collected from osteopontin-treated PC-3 cells were used aschemoattractrant (Fig. 5A and B). In separate experiments,endothelial cells were pretreated with anti-EP2 blocking antibodyand used for migration and invasion assays. Our data showed thatconditioned media of osteopontin-treated PC-3 cells significantlyenhanced endothelial cell migration and invasion, whereas block-ing the EP2 receptor in these cells with its antibody significantlysuppressed the endothelial cell migration and invasion (Fig. 5A andB). The enhanced migration of endothelial cells was furtherconfirmed by wound assay under the same experimental condition(Fig. 5C). Our results indicated that osteopontin-induced tumorcell–derived PGE2 enhanced migration and invasion of endothelialcells through an EP2-mediated paracrine mechanism.Development of PC-3 xenograft model to study the role of

COX-2 in osteopontin-induced tumor progression. Our in vitroresults prompted us to study the role of COX-2 and PGE2 inosteopontin-induced mice xenograft tumor progression. Accord-ingly, PC-3 cells (5 � 105) were mixed with Matrigel and theninjected s.c. into the dorsal flanks of the male athymic nude mice.In other experiments, cells were treated with osteopontin alone oralong with anti-EP2 blocking antibody and then injected into thenude mice. The osteopontin alone or mixture of osteopontin andanti-EP2 antibody was also injected to the tumor sites twice a weekfor up to 4 weeks. In another experiments, celecoxib (1,500 ppm)was supplemented to the normal diet of the osteopontin-injectedanimals. All the mice were kept under pathogen-free conditions.

Figure 6A (a-d) showed typical photographs of tumors grown innude mice. Blood was collected; mice were sacrificed by cervicaldislocation; and tumors were excised, weighed, and measured(Table 1). Tumor samples were used for histopathologic andimmunohistochemical studies according to standard procedure.The histopathologic analysis by H&E staining is summarized inTable 2, and these data clearly indicated that osteopontin-inducedtumorigenicity in nude mice was suppressed significantly by anti-EP2 blocking antibody or by supplementation of celecoxib in micediet (Table 2; Fig. 6A, e-h). Moreover, the immunohistochemicalstudies showed that the enhanced expressions of COX-2 and MMP-2 were detected in mice treated with osteopontin (Fig. 6A, i-p). Theexpression of vWF (an endothelial cell–specific marker) was alsodetected (Fig. 6A, q-t). The plasma PGE2 level was also higher inosteopontin injected mice (Fig. 6B). Taken together, our in vivoxenograft study showed that osteopontin induces prostate tumorgrowth and angiogenesis in nude mice that is correlated withup-regulation of COX-2 expression, MMP-2 activation, and PGE2

secretion. Inhibition of COX-2 or blocking of EP2 receptor signi-ficantly suppressed osteopontin-induced angiogenesis and tumorprogression in nude mice and further showed that COX-2plays an important role in osteopontin-induced prostate cancerprogression.Expressions of osteopontin, COX-2, MMP-2, and NF-KB, p65

localization, and their correlation with human prostatecancer progression and angiogenesis in different pathologicgrades. To correlate the in vitro data and in vivo mouse modelresults with human clinical specimens, human prostate cancertissues were collected from a local hospital with informed consent.The tumor samples were stained with H&E, and the grades of thesesamples were determined by Gleason grading system with the helpof expert histoplathologist (Fig. 7A, a-c). Expression of osteopontin,COX-2, MMP-2, and vWF and cellular localization of p65were analyzed by immunohistochemistry using their specific

Figure 5. Osteopontin (OPN ) stimulatesEP2-mediated endothelial cell migration,invasion, and wound healing via paracrinemechanism. A, human endothelial(EA.hy-926) cells were added in the upperchamber. The conditioned medium (CM)collected from PC-3 cells treated withosteopontin was used as chemoattractantin the lower chamber. In separateexperiments, EA.hy-926 cells were treatedwith anti-EP2 blocking antibody and usedfor migration assay. After 12 hours,migrated cells were counted as describedearlier. B and C, experiments parallelingthose of (A ) but measuring invasionand wound healing rather than migration.Columns, means of three determinations;bars, SE.

Cancer Research


antibodies (Fig. 7A, d-r). These results indicated the higher levels ofosteopontin and COX-2 in malignant tumors than normal and PIN,and that further correlated with the enhanced MMP-2 expressionand neovascularization (vWF expression). Moreover, there wassignificant nuclear translocation of p65 in the malignant tumors.All these data correlate with our in vitro and mice xenograftstudies.

Discussion

The interaction between osteopontin and prostate cancer cellsare likely to be the key determinants in regulating tumorprogression and metastatic phenotype of human prostate carcino-ma and further indicates that osteopontin may be an importantmediator of prostate cancer progression (40). Recent strategies intherapeutics of prostate cancers with inhibitors that target COX-2(41) and its enhanced expression in prostate carcinoma suggestthat understanding the molecular mechanism underlying enhancedCOX-2 expression may help in developing the novel therapeuticapproach in prostate cancer treatment. The experimental workpresented in the study showed for the first time the molecularmechanism that underlies osteopontin-induced COX-2 expressionand its potential role in regulating in vitro cell motility and

invasion of prostate cancer, which ultimately modulates in vivotumor growth and angiogenesis. Silencing of osteopontin in pros-tate cancer cells results drastic reduction, whereas overexpressionof osteopontin significantly increased COX-2 level, suggesting thespecificity of osteopontin on COX-2 expression. Moreover, we findthat COX-2 plays an important role in osteopontin-induced PGE2

production and MMP-2 activation that ultimately regulates tumorprogression.

Figure 6. COX-2 and PGE2 play crucial roles in osteopontin (OPN )–induced tumorigenesis and angiogenesis in nude mice. A, photographs of athymic mice showing4-week-old xenograft tumor growth by PC-3 cells; inset, excised tumors with respective mice (a-d). Experimental details were described in Materials and Methods.Representative H&E-stained sections from PC-3 xenograft tumors and the characteristics of these tumors were analyzed (e-h). The expression of COX-2 (i-l ),MMP-2 (m-p ), and neovascularization (vWF expression; q-t ) were visualized by immunofluorescence study using their specific antibodies. COX-2, MMP-2, and vWFwere stained with FITC-conjugated IgG (green ). B, levels of PGE2 production in these mice were determined from mice plasma by enzyme immunoassay. Threemice were used in each set of experiments.

Table 1. Effect of COX-2 inhibition and EP2 receptorblocking on osteopontin-induced prostate tumor growth innude mice

No. mice Treatments Tumor wt

(fold changes)

3 Control 1.00 F 0.13

3 Osteopontin 2.83 F 0.17

3 Osteopontin + celecoxib (1,500 ppm) 0.52 F 0.143 Osteopontin + EP2 antibody (20 Ag) 0.12 F 0.10



In prostate cancer, PKCa plays a key role in the regulationof downstream oncogenic molecules (34). Recent reports alsorevealed that activation of PKCa is required for the survival andgrowth of androgen-independent human prostate cancer cells(41–43). However, it is not well established how osteopontin regu-lates PKCa activation and PKCa-dependent downstream signalingevents in prostrate carcinoma. In this study, we provide evidences

that PKCa plays an important role in osteopontin-induced c-Src/NIK–mediated IKKa/h-dependent NF-nB phosphorylation, whichultimately controls COX-2 expression. Recently, we have shown theinvolvement of the NIK/IKK/NF-nB signaling pathway in osteo-pontin-induced expression of downstream effecter molecules thatregulate melanoma and breast tumor progression (21, 27). In thisstudy, we report that osteopontin regulates PKCa/c-Src–mediated

Table 2. Characteristics of the xenograft tumors from experimental mice

Tumor characteristics Control Osteopontin (0.5 Amol/L) Osteopontin (0.5 Amol/L) +

celecoxib (1,500 ppm)

Osteopontin (0.5 Amol/L) +

EP2 antibody (20 Ag)

Tumor infiltration Moderate Very high Moderate to poor Poor

Vessel formation Poor High Moderate to poor NegligibleMitotic features/hpf 4-6 12-16 2-4 1-3

Tumor giant cells Less Plenty Scanty Very less

Nuclear polymorphism Moderate nuclearsize variation

Marked nuclear size variation Less nuclear size variation Small regular uniformnucleus

Figure 7. Expressions of osteopontin (OPN ), COX-2, MMP-2, and NF-nB; p65 localization; and their correlation with human prostate cancer progression andangiogenesis in different pathological grades. A, prostate tumor specimens were collected from a local hospital with informed consent. The gradations of thesespecimens were done according to the Gleason grading system by H&E staining (a-c ). The levels of osteopontin (d-f ), COX-2 (j-l ), vWF (m-o ), and MMP-2 (p-r)expression and cellular localization of p65 (g-i ) were detected by immunohistochemical studies using their specific antibodies. Osteopontin, vWF, and MMP-2were stained with FITC (green ), whereas COX-2 and p65 (NF-nB) were stained with TRITC (red)–conjugated IgG, and the nucleus was counterstained with4¶,6-diamidino-2-phenylindole (blue ). B, schematic representation of osteopontin-induced PKCa/c-Src/IKK/NF-nB–mediated COX-2 expression leading to enhancedPGE2 production and MMP-2 activation that further induces tumorigenesis and angiogenesis via autocrine and paracrine mechanisms.

Cancer Research


IKK activation followed by NF-nB phosphorylation. Moreover, theresults showed that overexpression of super-repressor form of InBasignificantly suppressed osteopontin-induced COX-2 expression,suggesting that NF-nB act as the key transcription factor inregulation of COX-2 expression and further showed the dispens-ability of the PKCa/c-Src/IKK/NF-nB signaling cascade in osteo-pontin-induced COX-2 expression.

It is well established that increased level of PGE2 production andMMP-2 activation is associated with enhanced expression of COX-2in many cancers (44, 45). Administration of COX-2 inhibitors hasbeen reported to suppress PGE2 production and MMP-2 activationin various cancers (46–48). Recent reports also indicated that PGE2

could also induce MMP-2 activation in cancer cells (49). Our datashowed that inhibition of upstream signaling pathway or treatmentwith COX-2 inhibitor (celecoxib) significantly suppressed osteo-pontin-induced PGE2 production and MMP-2 activation.

Recent findings indicated that COX-2-specific inhibitors signif-icantly suppressed tumor growth in prostate cancer (50).Interestingly, our studies showed that NSAID celecoxib andNS398 (COX-2-selective inhibitor) significantly suppressed osteo-pontin-induced PC-3 cells migration and invasion. The in vivoxenograft tumor experiment indicated that mice fed with celecoxibshowed significant reduction in osteopontin-induced tumorgrowth, and this study suggested that COX-2 could be an effectivetarget in caner therapy.

Tumor cell–derived PGE2 can contribute tumor progressionby regulating the cell motility/invasiveness and inducing angio-genesis (8–10). Among PGE2 receptors, EP2 is expressed in bothprostate cancer and endothelial cells (39, 51). Moreover, thedevelopment of selective antagonists against the EP2 receptor has

the potential to improve antitumor activity, and EP2 receptorantagonist may be more specific than the use of COX-2 inhibitors(52). Sung et al. showed that tumors from wt EP2 mice producedmore blood vessels than those of knockout mice (53). In our study,we showed that the blocking the EP2 receptor by specific blockingantibody suppressed osteopontin-induced prostate cancer (PC-3)and endothelial cell motility and invasiveness. Furthermore, wealso find that administration of anti-EP2 blocking antibodysignificantly suppressed osteopontin-induced mice xenografttumor progression.

Our study showed the detailed molecular mechanism by whichosteopontin induces cell migration, invasion, and tumor progres-sion through induction of COX-2 expression and PGE2 production(Fig. 7B). Our results further warrant that the mechanism shown inthe mouse model underlies the human pathology and a clearunderstanding of osteopontin and COX-2 regulation couldilluminate cellular changes that accompany prostate cancerprogression and may facilitate the development of novel thera-peutic approaches to suppress osteopontin-regulated PKCa/IKK/NF-nB–mediated COX-2 expression and thereby controlling tumorprogression and angiogenesis.

Acknowledgments

Received 2/20/2006; accepted 5/2/2006.Grant support: National Bioscience Award Fund for Career Development (G.C.

Kundu); Department of Biotechnology, Government of India; and Council of Scientificand Industrial Research, Government of India (S. Jain).

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Ajay Gangshettiwar for histopathologic studies.

References1. McCarty MF. Targeting multiple signaling pathways asa strategy for managing prostate cancer: multifocalsignal modulation therapy. Integr Cancer Ther 2004;3:349–80.

2. Wallace JM. Nutritional and botanical modulation ofthe inflammatory cascade—eicosanoids, cyclooxyge-nases, and lipoxygenases—as an adjunct in cancertherapy. Integr Cancer Ther 2002;1:7–37.

3. DeWitt D, Smith WL. Yes, but do they still getheadaches? Biochem Cell 1995;83:345–8.

4. DeWitt DL. Prostaglandin endoperoxide synthase:regulation of enzyme expression. Biochim Biophys Acta1991;1083:121–34.

5. Smith WL, Meade EA, DeWitt DL. Pharmacology ofprostaglandin endoperoxide synthase isozymes-1 and -2.Ann N Y Acad Sci 1994;714:136–42.

6. Chandrasekharan NV, Dai H, Roos KL, et al. COX-3, acyclooxygenase-1 variant inhibited by acetaminophenand other analgesic/antipyretic drugs: cloning, struc-ture, and expression. Proc Natl Acad Sci U S A 2002;99:13926–31.

7. Chaudry AA, Wahle KW, McClinton S, Moffat LE.Arachidonic acid metabolism in benign and malignantprostatic tissue in vitro : effects of fatty acids andcyclooxygenase inhibitors. Int J Cancer 1994;57:176–80.

8. Wang D, DuBois RN. Cyclooxygenase 2-derivedprostaglandin E2 regulates the angiogenic switch. ProcNatl Acad Sci U S A 2004;101:415–6.

9. Houchen CW, Sturmoski MA, Anant S, Breyer RM,Stenson WF. Prosurvival and antiapoptotic effects ofPGE2 in radiation injury are mediated by EP2 receptorin intestine. Am J Physiol Gastrointest Liver Physiol2003;284:490–8.

10. Gately S, Li WW. Multiple roles of COX-2 in tumor

angiogenesis: a target for antiangiogenic therapy. SeminOncol 2004;31:2–11.

11. Sales KJ, Katz AA, Davis M, et al. Cyclooxygenase-2expression and prostaglandin E (2) synthesis are up-regulated in carcinomas of the cervix: a possibleautocrine/paracrine regulation of neoplastic cell func-tion via EP2/EP4 receptors. J Clin Endocrinol Metab2001;86:2243–9.

12. Rangaswami H, Bulbule A, Kundu GC. Osteopontin:role in cell signaling and cancer progression. Trends CellBiol 2006;16:79–87.

13. Rittling SR, Chambers AF. Role of osteopontin intumor progression. Br J Cancer 2004;90:1877–81.

14. Denhardt DT, Guo X. Osteopontin: a protein withdiverse functions. FASEB J 1993;7:1475–82.

15. Weber GF, Ashkar S, Glimcher MJ, Cantor H.Receptor-ligand interaction between CD44 and osteo-pontin (Eta-1). Science 1996;271:509–12.

16. Panda D, Kundu GC, Lee BI, et al. Potential roles ofosteopontin and alphaVbeta3 integrin in the develop-ment of coronary artery restenosis after angioplasty.Proc Natl Acad Sci U S A 1997;94:9308–13.

17. Liaw L, Birk DE, Ballas CB, Whitsitt JS, Davidson JM,Hogan BL. Altered wound healing in mice lacking afunctional osteopontin gene (spp1). J Clin Invest 1998;101:1468–78.

18. Sodek J, Ganss B, McKee MD. Osteopontin. Crit RevOral Biol Med 2000;11:279–303.

19. Angelucci A, Festuccia C, D’Andrea G, Teti A, BolognaM. Osteopontin modulates prostate carcinoma invasivecapacity through RGD-dependent upregulation of plas-minogen activators. Biol Chem 2002;383:229–34.

20. Philip S, Bulbule A, Kundu GC. Osteopontin stim-ulates tumor growth and activation of promatrixmetalloproteinase-2 through nuclear factor-kappa B-mediated induction of membrane type 1 matrix metal-

loproteinase in murine melanoma cells. J Biol Chem2001;276:44926–35.

21. Das R, Mahabeleshwar GH, Kundu GC. Osteopontinstimulates cell motility and nuclear factor kappa B-mediated secretion of urokinase type plasminogenactivator through phosphatidylinositol 3-kinase/Aktsignaling pathways in breast cancer cells. J Biol Chem2003;278:28593–606.

22. Aprikian AG, Tremblay L, Han K, Chevalier S.Bombesin stimulates the motility of human prostate-carcinoma cells through tyrosine phosphorylation offocal adhesion kinase and of integrin-associated pro-teins. Int J Cancer 1997;72:498–504.

23. Huang WC, Chen JJ, Inoue H, Chen CC. Tyrosinephosphorylation of I-kappa B kinase alpha/beta byprotein kinase C-dependent c-Src activation is involvedin TNF-alpha-induced cyclooxygenase-2 expression.J Immunol 2003;170:4767–75.

24. Rodrigues S, Nguyen QD, Faivre S, et al. Activation ofcellular invasion by trefoil peptides and src is mediated bycyclooxygenase- and thromboxane A2 receptor-depen-dent signaling pathways. FASEB J 2001;15:1517–28.

25. Baeuerle PA, Baltimore D. NF-kappa B: ten yearsafter. Cell 1996;87:13–20.

26. Malinin NL, Boldin MP, Kovalenko AV, Wallach D.MAP3K-related kinase involved in NF-kappaB inductionby TNF, CD95 and IL-1. Nature 1997;385:540–4.

27. Rangaswami H, Bulbule A, Kundu GC. Nuclear factor-inducing kinase plays a crucial role in osteopontin-induced MAPK/IkappaBalpha kinase-dependent nuclearfactor kappaB-mediated promatrix metalloproteinase-9activation. J Biol Chem 2004;279:38921–35.

28. Inoue H, Yokoyama C, Hara S, Tone Y, Tanabe T.Tanscriptional regulation of human prostaglandinendoperoxide synthase-2 gene by lipopolysaccharideand phorbol esters in vascular endothelial cells.



Involvement of both nuclear factor for interleukin-6expression site and cAMP response element. J BiolChem 1995;270:24965–71.

29. Inaba T, Sano H, Kawahito Y, et al. Induction ofcyclooxygenase-2 in monocyte/macrophage by mucinssecreted from colon cancer cells. Proc Natl Acad SciU S A 2003;100:2736–41.

30. Zhang P, Ostrander JH, Faivre EJ, Olsen A, Fitzsim-mons D, Lange CA. Regulated association of proteinkinase B/Akt with breast tumor kinase. J Biol Chem2005;280:1982–91.

31. Leopold JA, Walker J, Scribner AW, et al. Glucose-6-phosphate dehydrogenase modulates vascular endothe-lial growth factor-mediated angiogenesis. J Biol Chem2003;278:32100–6.

32. Zweifel BS, Davis TW, Ornberg RL, Masferrer JL.Direct evidence for a role of cyclooxygenase-2-derivedprostaglandin E2 in human head and neck xenografttumors. Cancer Res 2002;62:6706–11.

33. Chakraborty G, Rangaswami H, Jain S, Kundu GC.Hypoxia regulates crosstalk between Syk and Lckleading to breast cancer progression and angiogenesis.J Biol Chem 2006;281:11322–31.

34. Stewart JR, O’Brian CA. Protein Kinase C-alphamediates epidermal growth factor receptor transactiva-tion in human prostate cancer cells. Mol Cancer Ther2005;4:726–32.

35. Ghosh S, Karin M. Missing pieces in the NF-kappa Bpuzzle. Cell 2002;109:S81–96.

36. Kosaka T, Miyata A, Ihara H, et al. Characteriza-tion of the human gene (PTGS2) encoding prosta-glandin-endoperoxide synthase 2. Eur J Biochem 1994;221:889–97.

37. Timoshenko AV, Xu G, Chakrabarti S, Lala PK,

Chakraborty C. Role of prostaglandin E 2 receptors inmigration of murine and human breast cancer cells. ExpCell Res 2003;289:265–74.

38. Salcedo R, Zhang X, Young HA, et al. Angiogeniceffects of PGE 2 are mediated by up-regulation ofCXCR4 on human microvascular endothelial cells. Blood2003;102:1966–77.

39. Cattaneo MG, Pola S, Deho V, Sanguini AM, VicentiniLM. Alprostadil suppresses angiogenesis in vitro andin vivo in the murine matrigel plug assay. Br JPharmacol 2003;138:377–85.

40. Thalmann GN, Sikes RA, Devoll RE, et al. Osteopon-tin: possible role in prostate cancer progression. ClinCancer Res 1999;5:2271–7.

41. Karamouzis MV, Papavassiliou AG. COX-2 inhibi-tion in cancer therapeutics: a field of controversy ora magic bullet? Expert Opin Investig Drugs 2004;13:359–72.

42. Koren R, Ben Meir D, Langzam L, et al. Ex-pression of protein kinase C isoenzymes in benignhyperplasia and carcinoma of prostate. Oncol Rep2004;11:321–6.

43. O’Brian CA. Protein kinase C-alpha: a novel target forthe therapy of androgen-independent prostate cancer?Oncol Rep 1998;5:305–9.

44. Mayoral R, Fernandez-Martinez A, Bosca L, Martin-Sanz P. Prostaglandin E2 promotes migration andadhesion in hepatocellular carcinoma cells. Carcinogen-esis 2005;26:753–61.

45. Dohadwala M, Batra RK, Luo J, et al. Autocrine/paracrine prostaglandin E2 production by non-small celllung cancer cells regulates matrix metalloproteinase-2and CD44 in cyclooxygenase-2-dependent invasion.J Biol Chem 2002;277:50828–33.

46. Pan MR, Chuang LY, Hung WC. Non-steroidal anti-inflammatory drugs inhibit matrix metalloproteinase-2expression via repression of transcription in lung cancercells. FEBS Lett 2001;508:365–8.

47. Nagatsuka I, Yamada N, Shimizu S, et al. Inhibitoryeffect of a selective cyclooxygenase-2 inhibitor onliver metastasis of colon cancer. Int J Cancer 2002;100:515–9.

48. Attiga FA, Fernandez PM, Weeraratna AT, ManyakMJ, Patierno SR. Inhibitors of prostaglandin synthesisinhibit human prostate tumor cell invasiveness andreduce the release of matrix metalloproteinases. CancerRes 2000;60:4629–37.

49. Ito H, Duxbury M, Benoit E, et al. Prostaglandin E2enhances pancreatic cancer invasiveness through anEts-1-dependent induction of matrix metalloproteinase-2. Cancer Res 2004;64:7439–46.

50. Gupta S, Adhami VM, Subbarayan M, et al. Sup-pression of prostate carcinogenesis by dietary supple-mentation of celecoxib in transgenic adenocarcinomaof the mouse prostate model. Cancer Res 2004;64:3334–43.

51. Chen Y, Hughes-Fulford M. Prostaglandin E2 and theprotein kinase A pathway mediate arachidonic acidinduction of c-fos in human prostate cancer cells. Br JCancer 2000;82:2000–6.

52. Pavlovic S, Du B, Sakamoto K, et al. Targeting PGE 2receptors as an alternative strategy to block COX-2-dependent extracellular matrix-induced MMP-9 expres-sion by macrophages. J Biol Chem 2006;281:3321–8.

53. Sung YM, He G, Fischer SM. Lack of expression of theEP2 but not EP3 receptor for prostaglandin E2 results insuppression of skin tumor development. Cancer Res2005;65:9304–11.

Cancer Research


Prostaglandin E2 Regulates Tumor Angiogenesis in Prostate Cancer

Shalini Jain, Goutam Chakraborty, Remya Raja, Smita Kale, and Gopal C. Kundu


Abstract

In cancer management, the cyclooxygenase (COX)–targetedapproach has shown great promise in anticancer therapeutics.However, the use of COX-2 inhibitors has side effects andhealth hazards; thus, targeting its major metabolite prosta-glandin E2 (PGE2)–mediated signaling pathway might be arational approach for the next generation of cancer manage-ment. Recent studies on several in vitro and in vivo modelshave revealed that elevated expression of COX-2 correlateswith prostate tumor growth and angiogenesis. In this study,we have shown the in-depth molecular mechanism and thePGE2 activation of the epidermal growth factor receptor andB3 integrin through E prostanoid 2 (EP2)–mediated andEP4-mediated pathways, which lead to activator protein-1(AP-1) activation. Moreover, PGE2 also induces activatingtranscription factor-4 (ATF-4) activation and stimulates cross-talk between ATF-4 and AP-1, which is unidirectional towardAP-1, which leads to the increased expressions of urokinase-type plasminogen activator and vascular endothelial growthfactor and, eventually, regulates prostate tumor cell motility.In vivo Matrigel angiogenesis assay data revealed that PGE2

induces angiogenesis through EP2 and EP4. Human prostatecancer specimen analysis also supported our in vitro andin vivo studies. Our data suggest that targeting PGE2 signalingpathway (i.e., blocking EP2 and EP4 receptors) might be arational therapeutic approach for overcoming the side effectsof COX-2 inhibitors and that this might be a novel strategy forthe next generation of prostate cancer management. [CancerRes 2008;68(19):7750–9]

Introduction

Treatment of cancer by chemotherapeutic agents is consideredone of the most effective approaches in cancer management inrecent times. Earlier reports have depicted that reduced apoptosis,increased neovascularization, and immunosuppression are some ofthe known consequences of cyclooxygenase-2 (COX-2) overexpres-sion, and each effect could have an important role in tumorprogression and angiogenesis (1). Several selective and nonselectiveCOX-2 inhibitors have been in use for the treatment of differentcancers, but many questions have arisen regarding their side effects(2). Various studies have shown the correlation between COX-2overexpression and enhanced production of prostaglandin E2(PGE2) by cancer cells (3). It has been reported that the rate ofPGE2 conversion from arachidonic acid is almost 10-fold higher inmalignant prostatic tissues than in benign prostatic tissues (4).

Thus, the concerns regarding the safety of these COX-2 inhibitors,as well as the identification of the more effective therapeuticagents, prompted us to understand the downstream signalingevents regulated by PGE2 in prostate cancer, which might help todevelop new therapeutic approach in the treatment of prostatecancer.

PGE2 interacts with the E prostanoid (EP) family of receptors,which consist four different subtypes (EP1–EP4). The enhancedexpressions of EP2 and EP4 receptors have been shown in prostatecancer, as well as in endothelial cells (5, 6). In this study, we haveexamined the role of PGE2-mediated signaling during prostatecancer progression and suggested that blocking the interactionbetween PGE2 and its receptors, rather than global prostaglandinsynthesis by using specific COX-2 inhibitors, might circumventsome of the adverse side effects. Recently, we have shown that thechemokine-like protein, osteopontin, induces COX-2–dependentPGE2-mediated prostate cancer progression (7). However, themolecular mechanism by which PGE2 directly regulates prostatecancer progression and angiogenesis is not well defined.

Previous studies have shown that PGE2 augments cyclic AMP(cAMP) production (8), increases cellular growth, and regulatesdifferentiated cell functions by promoting the activation of cAMP-dependent protein kinase A (PKA). The PKA-mediated phosphor-ylation of cAMP-responsive element binding protein (CREB) andregulation of transcription via interaction between cAMP-responseelements with CREB are considered as the major pathways thatalter gene expression in cancer cells (9, 10). Earlier studies haverevealed that activating transcription factor 4 (ATF-4; also calledCREB-2) regulates the expression of genes involved in oxidativestress, amino acid synthesis, differentiation, metastasis, andangiogenesis (11). It has been reported that the expression ofATF-4 is induced by various external stimuli in cancer microen-vironment and regulates various processes that control cancerprogression (11), but the function of ATF-4 in prostate cancerprogression remains unknown.

The overexpression of proteases often correlates with theenhanced tumor cell invasion and metastasis by virtue ofdegradation of extracellular matrix and basement membranes inalmost all malignancies, including prostate cancer (12, 13).Urokinase-type plasminogen activator (uPA), a protease, plays animportant role in tumor cell invasion and metastasis (14).Increased expressions of uPA and vascular endothelial growthfactor (VEGF) have been reported in malignancies of variousorgans including prostate (14, 15), and the increased expression ofthese molecules is associated with an enhanced metastatic andangiogenic potential and poor survival of patients (16). Earlier datahave shown that the response elements for activator protein 1(AP-1) and ATF-4 are present in the promoter region of uPA andVEGF (17–20). Although it has been reported that PGE2 plays acrucial role in VEGF production in prostate cancer cells (21), themolecular mechanism by which PGE2 regulates ATF-4/AP-1–mediated uPA and VEGF expressions, which lead to prostatetumor cell motility and in vivo angiogenesis, remains unknown.

Note: Supplementary data for this article are available at Cancer Research Online(http://cancerres.aacrjournals.org/).

Requests for reprints: Gopal C. Kundu, National Center for Cell Science, NCCSComplex, Pune 411 007, India. Phone: 91-20-25708103; Fax: 91-20-25692259; E-mail:[email protected].


Cancer Res 2008; 68: (19). October 1, 2008 7750 www.aacrjournals.org

Research Article

In this study, we have shown that PGE2 triggers mitogen-activated protein kinase (MAPK)/extracellular signal-regulatedkinase (ERK) kinase (MEK)/ERK1/2 signaling through activationof epidermal growth factor receptor (EGFR) and augmentsexpression and activation of h3 integrin in prostate cancer cells.Moreover, we have shown that PGE2 induces the activation of ATF-4 and AP-1 via EGFR-MEK-ERK1/2 or h3 integrin–mediatedpathway, which ultimately leads to the increased expressions ofuPA and VEGF. Furthermore, we have observed that PGE2 regulatesendothelial cell motility and in vivo angiogenesis. Analysis ofhuman prostate clinical samples showed that the expressionprofiles of EP2 and EP4 receptors correlated with levels of AP-1,ATF-4, uPA, and VEGF. These data suggested that, at least in part,PGE2 plays a crucial role in the oncogenesis and angiogenesis ofprostate cancer. Thus, targeting PGE2 receptor–mediated signalingmight be a potential approach for the improved prostate cancertherapeutics.


Cell culture and transfection. Human prostate cancer cell lines (PC-3,

DU-145, and LNCaP) were obtained from American Type Culture Collection.

Human umbilical vein endothelial cell (HUVEC) was purchased from

Cambrex. COX-2 cDNA (Dr. Stephen Prescott, University of Utah), wild-type(wt) and dominant-negative (dn) ATF-4 (wt, pEF/mATF-4; dn, pEF/mATF-

4M; Dr. Javed Alam, Yale University School of Medicine), and A Fos (Dr.

Charles Vinson, National Cancer Institute) were transfected in PC-3 cells

using Lipofectamine 2000.Small interfering RNA. PC-3 cells were transfected with small

interfering RNA (siRNA) that specifically targets COX-2 (COX-2 siRNA,

Santa Cruz Biotechnology), EP2 (ON-TARGET plus SMARTpool PTGER2; L-005712-00), EP4 (siGENOME SMARTpool PTGER4; M-005714-00), human

integrin h3 (siGENOME SMARTpool ITGB3; M-004124-02), and control

siRNA (siGENOME nontargeting siRNA; D-001206-14-05 and ON-TARGET

plus nontargeting pool; D-001810-10-05; Dharmacon) according to themanufacturer’s instructions.

Western blot and EMSA. The Western blot and EMSA experiments were

performed as described earlier (7, 22).

Immunofluorescence. Immunofluorescence studies were performedusing specific antibodies as described earlier (22, 23).

Flow cytometry. Flow cytometry experiments were performed as

described (24).Reverse transcription–PCR. Total RNA was isolated from PC-3 cells and

analyzed by reverse transcription–PCR (RT-PCR). The following primers

were used: uPA sense, 5¶-CAC GCA AGG GGA GAT GAA-3¶; uPA antisense,

5¶-ACA GCA TTT TGG TGG TGA CTT-3¶; VEGF sense, 5¶-CCT CCG AAA CCATGA ACT TT-3¶; VEGF antisense, 5¶-AGA GAT CTG GTT CCC GAA AC-3¶;h-actin sense, 5¶-GGC ATC CTC ACC CTG AAG TA-3¶; h-actin antisense,

5¶-GGG GTG TTG AAG GTC TCA AA-3¶. The amplified cDNA fragments were

analyzed by 1.5% agarose gel electrophoresis.Cell migration and comigration assay. The migration and comigration

experiments were performed as described (22). Briefly, PC-3 cells, either

alone or individually transfected with dn ATF-4, dn c-Jun, and A-Fos orpretreated with PKA inhibitor peptide, were added to the upper chamber of

the Boyden chamber. PGE2 was added in the upper chamber. For

comigration assay, PC-3 cells, either alone or transfected with COX-2 cDNA

or COX-2 siRNA, were used in the lower chamber. Endothelial (HUVEC)cells, either alone or pretreated with EP2 (AH6809, Sigma) or EP4 (AH23848,

Sigma) receptor antagonist, were used in the upper chamber. The migrated

endothelial cells to the reverse side of the upper chamber were fixed and

stained with Giemsa and counted in three high-power fields under aninverted microscope (Nikon). Data are represented as the average of three

counts F SE.

Wound assay. Wound assays were performed using PC-3 and endothelial

cells as described earlier (7). Wounds with a constant diameter were made.

PC-3 cells were treated with PGE2 alone or pretreated with EP2 or EP4receptor antagonist or PKA inhibitor peptide, or transfected with dn ATF-4,

dn c-Jun, A-Fos and then treated with PGE2. In separate experiments,

endothelial cells were treated with PGE2 alone or pretreated with EP2 or

EP4 receptor antagonist and then treated with PGE2. After 12 h, woundphotographs were taken through a microscope (Nikon).

In vivo Matrigel plug assay. In vivo Matrigel plug angiogenesis assay

was carried out, as described previously (22). Briefly, Matrigel, either alone

or along with PGE2, was injected s.c. into the ventral groin region of maleathymic NMRI (nu/nu) mice. In separate experiments, PGE2 containing

Matrigel was mixed with EP2 or EP4 antagonist (30 Amol/L) and injected

into the mice. In another experiment, conditioned medium of PC-3 cells,

either nontransfected or transfected with COX-2 cDNA, was mixed withMatrigel and injected into the mice. After 21 d, mice were sacrificed,

dissected out, and photographed. The Matrigel plugs were excised and used

for immunohistochemistry. The paraffin sections were immunostained withanti-vWF (Chemicon), anti-CD31 (Chemicon), anti–phosphorylated p65,

nuclear factor-nB (NF-nB; Cell Signaling Technology), and anti–phosphor-

ylated Akt (Santa Cruz) antibodies and visualized under confocal

microscope (Ziess).Human prostate cancer specimen analysis. Specimens of different

Gleason grades and normal tissues of prostate were collected from a local

hospital with informed consent and analyzed by immunohistochemistry as

described (7). The expression profiles of EP2, EP4, ATF-4, c-Jun, c-Fos, uPA,and VEGF were detected by immunohistochemistry using their specific

antibodies. Five specimens from each group [normal, prostatic intra-

epithelial neoplasia (PIN), and malignant] were analyzed.Statistical analysis. The data reported in cell migration, comigration,

in vivo Matrigel plug angiogenesis, and the clinical specimen analysis are

expressed as mean F SE. Statistical differences were determined by

Student’s t test. A P value of <0.05 was considered significant. All bandswere analyzed densitometrically (Kodak Digital Science), and fold changes

were calculated. The in vivo angiogenesis and clinical specimen data were

quantified using the Image Pro Plus 6.0 Software (Nikon).

Results

PGE2 augments EP2/EP4-mediated EGFR/MAPK and B3integrin activation in prostate cancer cells. To examine theeffect of PGE2 on EGFR, MEK, ERK1/2, and h3 integrinphosphorylation, serum-starved PC-3 cells were treated withPGE2 in a dose (0–1.0 Amol/L)–dependent and time (0–60minutes)–dependent manner. The levels of phosphorylation ofthese signaling molecules were analyzed by Western blot usingtheir phosphorylated-specific antibodies. The data indicatedthat PGE2 induces phosphorylation of EGFR, MEK, ERK1/2, andh3 integrin, and maximum phosphorylations were observedbetween 10 and 15 minutes (Fig. 1A) with 0.5 Amol/L ofPGE2 (Supplementary Fig. S1A). Moreover, the effect of PGE2 onthe activation of these signaling molecules (EGFR, MEK, ERK, andh3 integrin) was examined in other prostate cancer (DU-145 andLNCaP) cells. The data showed significant phosphorylations ofthese molecules in DU-145 compared with LNCaP cells in responseto PGE2 (Supplementary Fig. S1B). Previous reports have shownthat PC-3 cells express higher levels of EP2 and EP4 receptors (21).Therefore, to examine the involvement of EP2 and EP4 receptors inPGE2-induced EGFR and h3 integrin phosphorylation, PC-3 cellswere pretreated with EP2 (AH6809) or EP4 (AH23848) receptorantagonist in a dose-dependent manner (0–30 Amol/L) for 1 hourand then treated with PGE2, and the levels of phosphorylated EGFRand phosphorylated h3 integrin were analyzed by Western blot.AH6809 or AH23848 at 30 Amol/L concentration showed maximuminhibition of PGE2-induced EGFR and h3 integrin phosphorylation(Fig. 1B, I and II). To examine whether EP2 and EP4 receptor

PGE2-Dependent Tumor Progression and Angiogenesis

www.aacrjournals.org 7751 Cancer Res 2008; 68: (19). October 1, 2008

agonists mimic the effect of PGE2 and regulate the downstreammolecular events, PC-3 cells were treated with butaprost (EP2agonist) and PGE1 alcohol (EP4 agonist) and phosphorylated EGFRand phosphorylated h3 integrin were analyzed. The data showedthat both the agonists induce the phosphorylation of EGFR and h3integrin (Supplementary Fig. S2A and B). These data revealed thatPGE2 induces the phosphorylation of EGFR and h3 integrinthrough EP2 and EP4 receptors–mediated process. Recently, it hasbeen reported that PGE2 induces h1 integrin expression inhepatocellular carcinoma cells (25). To determine whether PGE2regulates the expression of h3 integrin in prostate cancer cells, PC-3 cells were treated with PGE2 for 12 hours and the expression ofh3 integrin was analyzed by flow cytometry (Fig. 1C). To determinethe roles of EP2 and EP4 receptors in PGE2-induced h3 integrinexpression, PC-3 cells were pretreated with AH6809 or AH23848and then treated with PGE2, and expression of h3 integrin wasanalyzed by immunofluorescence. The data showed that AH6809and AH23848 suppressed the PGE2-induced h3 integrin expression,

indicating that EP2 and EP4 receptors play crucial roles inregulating this process (Fig. 1D). These data suggested that PGE2does not only stimulate EGFR and h3 integrin phosphorylation butalso induces the expression of h3 integrin via EP2/EP4 receptor–mediated pathway.

EGFR and B3 integrin play crucial roles in PGE2-inducedAP-1 activation. Earlier studies have shown the role of AP-1 inprostate cancer progression (26, 27). Activation of AP-1 involves theincreased expression or activation of Jun and Fos proteins (28–30).To examine the effect of PGE2 on c-Fos and c-Jun expression/activation, PC-3 cells were treated with PGE2, and expressions of c-Fos and phosphorylation of c-Jun were analyzed by Western blotand immunofluorescence, whereas AP-1–DNA binding was per-formed by EMSA. The results revealed that PGE2 does not onlyaugment the expression of c-Fos and phosphorylation of c-Jun(Fig. 2A and Supplementary Fig. S3A) but also stimulates the AP-1–DNA binding (Supplementary Fig. S3B). Furthermore, to study therole of EGFR/MAPK or h3 integrin on PGE2-induced AP-1

Figure 1. PGE2 augments phosphorylation of EGFR, MEK, ERK, and h3 integrin in PC-3 cells. A, PC-3 cells were incubated with 0.5 Amol/L PGE2 for 0 to 60 min, andthe levels of p-EGFR, p-MEK, p-ERK, and p-h3 integrin were analyzed by Western blot using their specific antibodies. Total EGFR, MEK, ERK, and h3 integrinexpressions in the cells were used as loading controls. B, roles of EP2 and EP4 receptors in PGE2-induced phosphorylations of EGFR and h3 integrin. PC-3 cells werepretreated with either EP2 receptor antagonist (AH6809) or EP4 receptor antagonist (AH23848) in a dose-dependent manner (0–30 Amol/L) for 1 h and thentreated with PGE2, and the levels of phosphorylated EGFR and phosphorylated h3 integrin were analyzed by Western blot (I and II). C, PC-3 cells were treated withPGE2, and h3 integrin expression was analyzed by flow cytometry using anti–h3 integrin antibody. D, PC-3 cells were pretreated with AH6809 or AH23848 and thentreated with PGE2, and the level of h3 integrin was analyzed by immunofluorescence using anti-h3 integrin antibody followed by staining with Cy2-conjugated IgG(green ). Nuclei were stained with propidium iodide (PI, red). All figures are representation of three experiments. Fold changes were calculated.

Cancer Research


activation, PC-3 cells were pretreated with PD98059 (MEKinhibitor) or AG1478 (EGFR inhibitor) or transfected with h3integrin siRNA and expression of c-Fos and levels of thephosphorylated c-Jun were analyzed by Western blot. The datashowed that inhibition of EGFR-MAPK pathway or down-regulation of h3 integrin suppressed PGE2-induced c-Fos expres-sion and c-Jun phosphorylation, indicating that PGE2 triggersEGFR-MAPK and h3 integrin–mediated AP-1 activation (Fig. 2B,I and II). Altogether, these results suggested that EGFR and h3integrin play crucial roles in PGE2-induced AP-1 activation in PC-3cells.

PGE2 stimulates ATF-4–dependent AP-1 activation. Elevatedexpression of ATF-4 has been observed in various cancersassociated with enhanced malignancy (11). Recent findings haveshown that ATF-4 is also involved in the regulation of expression ofvarious oncogenic molecules and plays a crucial role in cancerprogression (11). Therefore, we have examined the expression ofATF-4 in PC-3, DU-145, and LNCaP cells by immunofluorescence.The results showed the significant level of ATF-4 expression,particularly in PC-3 and DU-145 cells (data not shown). Toinvestigate the role of PGE2 on ATF-4 activation, PC-3 cells weretreated with PGE2 and ATF-4 nuclear localization and DNA binding

Figure 2. PGE2 induces EP2/EP4-mediated EGFR/MAPK or h3 integrin–dependent c-Fos expression and c-Jun phosphorylation and enhances colocalization ofATF-4 with phosphorylated c-Jun. A, PC-3 cells were treated with 0.5 Amol/L of PGE2 for 0 to 150 min, and the levels of c-Fos and phosphorylated c-Jun were detectedby Western blot. Total c-Jun and actin were used as loading controls. B, PC-3 cells were pretreated with PD98059 (MEK inhibitor; 40 Amol/L) or AG1478 (EGFRinhibitor; 500 nmol/L) for 1 h and then treated with PGE2. c-Fos and phosphorylated c-Jun were analyzed by Western blot (I ). PC-3 cells were individually transfectedwith h3 siRNA or control siRNA and then treated with PGE2, and the levels of c-Fos and phosphorylated c-Jun were analyzed (II). C, PC-3 cells were treated with PGE2,and the colocalization of phosphorylated c-Jun (stained with Cy2; green ) and ATF-4 (stained with Cy3; red ) was analyzed by immunofluorescence. Nuclei werecounterstained with 4¶,6-diamidino-2-phenylindole (blue ). D, ATF-4 regulates PGE2-induced expression of c-Fos and phosphorylation of c-Jun. PC-3 cells weretransfected with wt and dn ATF-4 and then treated with PGE2, and the levels of c-Fos, phosphorylated c-Jun, and c-Jun were analyzed by Western blot. All figures arerepresentation of three experiments. Fold changes were calculated.



were determined by immunofluorescence and EMSA. The dataindicated that PGE2 induces nuclear localization and DNA bindingof ATF-4 (Supplementary Fig. S4A and B). Moreover, we haveobserved the enhanced nuclear colocalization of ATF-4 withphosphorylated c-Jun in response to PGE2 (Fig. 2C). Furthermore,to explore the cross-talk between ATF-4 and AP-1, PC-3 cells wereindividually transfected with wt or dn ATF-4, followed by treatmentwith PGE2. The levels of c-Fos and phosphorylated c-Junexpressions were analyzed by Western blot. The data indicatedthat wt ATF-4 enhances, whereas dn ATF-4 suppresses, PGE2-induced c-Fos and phosphorylated c-Jun expression (Fig. 2D).EMSA data further confirmed that ATF-4 regulates AP-1–DNAbinding in response to PGE2 (Supplementary Fig. S4C). However,wt and dn c-Jun or A-Fos had no effect on PGE2-induced ATF-4–DNA binding (data not shown), which further suggested that PGE2-regulated cross-talk between ATF-4 and AP-1 is unidirectionaltoward AP-1.

PGE2 induces EP2/EP4-mediated uPA and VEGF expressionsin prostate cancer cells. To examine the role of PGE2 on uPA andVEGF expressions, PC-3 cells were treated with PGE2 in a time-dependent (0–24 h) and dose-dependent (0–1.0 Amol/L) manner.

The levels of uPA and VEGF were analyzed by Western blot. Theresults indicated that PGE2 with 0.5 Amol/L concentration inducedmaximum expressions of uPA and VEGF at f16 hours (Fig. 3A andSupplementary Fig. S5A). Similarly, PGE2 at 0.5 Amol/L concentra-tion stimulated maximum uPA and VEGF expressions at mRNAlevels (Fig. 3B). The PGE2-induced uPA and VEGF expressions werealso detected in DU-145 and LNCaP cells (Supplementary Fig. S5B).The data indicated that PGE2 up-regulates uPA and VEGFexpressions, both at transcriptional and protein levels. Earlierreports have indicated that COX-2 regulates PGE2 production inprostate tumor cells (3). Therefore, to determine the role of tumor-derived PGE2 on uPA and VEGF expressions, PC-3 cells weretransfected with COX-2 cDNA or COX-2 siRNA (COX-2i) andexpressions of uPA and VEGF were detected by Western blot. Thedata showed that overexpression of COX-2 enhances, whereassilencing of COX-2 suppresses, uPA and VEGF expressions(Supplementary Fig. S5C), which further suggested that tumor-derived PGE2 is also involved in the regulation of uPA and VEGFexpressions in these cells. To delineate the roles of EP2 and EP4receptors in PGE2-induced uPA and VEGF expressions, PC-3 cellswere transfected with EP2 or EP4 siRNA (EP2i or EP4i) and then

Figure 3. PGE2 augments EP2/EP4 receptor–mediated ATF-4/AP-1–dependent uPA and VEGF expression in PC-3 cells. A, PC-3 cells were treated with 0.5 Amol/LPGE2 for 0 to 24 h, and the levels of uPA and VEGF were detected by Western blot using their specific antibodies. Actin was used as loading control. B, cellswere treated with PGE2 in a dose (0–1.0 Amol/L)–dependent manner, total RNA was isolated, and the levels of uPA and VEGF mRNA were detected by semiquantitativeRT-PCR. h-Actin was used as internal control. C, PC-3 cells were transfected with EP2 or EP4 specific siRNA (EP2 i or EP4 i) and then treated with PGE2, andthe levels of uPA, VEGF, EP2, and EP4 were analyzed by Western blot. Actin was used as loading control. D, roles of ATF-4 and AP-1 in PGE2-induced uPA and VEGFexpression. Cells were individually transfected with dn ATF-4, dn c-Jun, and A-Fos followed by treatment with PGE2, and the levels of uPA and VEGF were analyzed byWestern blot. All figures are representation of three experiments. Fold changes were calculated.

Cancer Research


treated with PGE2. The levels of uPA, VEGF, EP2, and EP4 wereanalyzed by Western blot. The data indicated that silencing of EP2or EP4 receptor suppresses PGE2-induced uPA and VEGFexpressions (Fig. 3C). Moreover, to examine the roles of ATF-4and AP-1 on PGE2-induced uPA and VEGF expressions, PC-3 cellswere transfected with dn ATF-4 or dn c-Jun or A-Fos cDNAconstruct and then treated with PGE2, and the levels of uPA andVEGF were analyzed by Western blot. The data revealed that dnATF-4, dn c-Jun, or A-Fos suppressed PGE2-induced uPA and VEGFexpressions (Fig. 3D), demonstrating the roles of AP-1 and ATF-4 inPGE2-induced uPA and VEGF expressions. Taken together, thesedata indicated that both exogenous and tumor-derived PGE2induced uPA and VEGF expressions via EP2 and EP4 receptors–mediated ATF-4–dependent and AP-1–dependent pathway.

ATF-4 and AP-1 regulate PGE2-induced prostate tumor cellmotility. It has been reported that nonsteroidal antiinflammatory

drugs and selective COX-2 inhibitors suppress invasiveness ofhuman prostate cancer cell lines, PC-3 and DU-145, and this effectcan be reversed by the addition of PGE2 (31). Although it has beenproposed that overexpression of COX-2/PGE2 may enhance theinvasive properties of tumors (3), leading to increase in tumor cellmigration, the molecular mechanism underlying this process is notwell defined. Therefore, to delineate the molecular mechanism ofPGE2-regulated tumor cell migration, PC-3 cells were eitherindividually transfected with dn ATF-4, dn c-Jun, and A-Fos orpretreated with AH6809, AH23848, and PKA inhibitor peptide andthen treated with PGE2, and wound migration assay wasperformed. These data showed that antagonists of EP2 and EP4receptors, PKA inhibitor peptide, dn ATF-4, dn c-Jun, and A Fossignificantly suppressed PGE2-induced tumor cell migration(Fig. 4A, I and II). The roles of these molecules in PGE2-mediatedPC-3 cell migration were further confirmed by migration assay

Figure 4. A, EP2 and EP4 receptors play crucial roles in PGE2-induced tumor cell motility. PC-3 cells were either individually transfected with dn ATF-4, dn c-Jun, andA-Fos or pretreated with AH6809 or AH23848 or PKA inhibitor peptide and then treated with PGE2, and wound migration assay was performed. Wound photographswere taken at 0 and 12 h (I and II ). B, PGE2 controls Akt and p65, NF-nB phosphorylation in endothelial (HUVEC) cells. HUVEC were treated with PGE2 for 0 to60 min, and the levels of phosphorylated Akt and phosphorylated p65 were detected by Western blot. Total Akt and p65 were used as control. C, HUVEC werepretreated with AH6809 or AH23848 and then treated with PGE2. The levels of phosphorylated Akt and phosphorylated p65 were analyzed by Western blot. Total Aktand p65 were used as control. Data represent three experiments exhibiting similar results. Fold changes were calculated.



using Boyden chamber (Supplementary Fig. S6A). Taken together,the results showed that PGE2 regulates ATF-4/AP-1–dependentprostate tumor cell motility through interaction with EP2 and EP4receptors.

PGE2 induces EP2/EP4-mediated Akt/NF-KB activation inendothelial cells, tumor-endothelial cell interaction, andangiogenesis. NF-nB regulates the expression of various factorsthat control endothelial and tumor cell motility and invasion (32).The serine/threonine protein kinase Akt is an important compo-nent in the migratory and prosurvival signaling pathways (33).Therefore, to examine the effect of PGE2 on the activation of NF-nBand Akt, endothelial cells (HUVEC) were treated with PGE2 in atime-dependent manner and the levels of phosphorylated Akt andphosphorylated p65 and NF-nB were analyzed by Western blot. Thedata showed that PGE2 induces the phosphorylations of Akt andp65 in these cells (Fig. 4B). To examine the role of EP2 or EP4receptor on PGE2-induced phosphorylation of Akt and p65, HUVECwere pretreated with AH6809 and AH23848 and then treated withPGE2, and the levels of phosphorylated Akt and phosphorylatedp65 were analyzed. The data showed that EP2 and EP4 receptorantagonists suppressed PGE2-induced phosphorylation of Akt andp65, suggesting that EP2 and EP4 play crucial roles in this process(Fig. 4C). To examine the roles of EP2 and EP4 on PGE2-mediatedendothelial cell motility, wound migration assay was performed.The data indicated that EP2, as well as EP4 receptor antagonists,suppressed PGE2-induced endothelial cell motility (SupplementaryFig. S6B).

Various studies have indicated that overexpression of COX-2 andPGE2 correlates with tumor angiogenesis (3, 7, 34). To determinethe role of tumor-derived PGE2 on endothelial cell motility, directcomigration assay was performed. PC-3 cells, either alone ortransfected with wt COX-2 cDNA or COX-2 siRNA, were used in thelower chamber, whereas HUVEC, either alone or pretreated withEP2 or EP4 receptor antagonist, were used in the upper chamber.In separate experiments, PGE2 was used in the lower chamber aspositive control. The endothelial cells migrated toward the reverseside of the upper chamber were stained with Giemsa, photo-graphed, counted, and represented in the form of a bar graph(Fig. 5A). The data revealed that overexpression of COX-2significantly enhanced, whereas silencing COX-2 or antagonists ofEP2 or EP4 receptor drastically suppressed, endothelial cell motilitytoward tumor cells, suggesting that tumor-derived PGE2 plays acrucial role in this process (Fig. 5A).

To examine the effect of PGE2 on in vivo tumor angiogenesis,Matrigel plug angiogenesis assay was performed. Accordingly,PGE2 was mixed with growth factor depleted Matrigel alone oralong with EP2 or EP4 receptor antagonists and Matrigel andinjected into the nude mice. After the termination of theexperiments, Matrigel plugs were photographed and analyzed byimmunohistochemistry using anti-CD31, anti-vWF, anti–phos-phorylated p65, and anti–phosphorylated Akt antibodies. Theresults showed that PGE2-induced angiogenesis was inhibited byEP2 or EP4 receptor antagonist (Fig. 5B). The expressions of vWFand CD31 (endothelial cell markers) and phosphorylations ofp65, NF-nB, and Akt were higher in PGE2-treated plugs comparedwith the plugs developed with EP2 and EP4 receptor antagonists(Fig. 5B). In other experiments, conditioned medium of PC-3 cells,either nontransfected or transfected with COX-2 cDNA, was mixedwith Matrigel and injected into the mice. The Matrigel plugsgenerated from the conditioned medium of COX-2 overexpressingPC-3 cells showed enhanced tumor angiogenesis compared with

the conditioned medium of PC-3 cells alone, suggesting thattumor-derived PGE2 plays a crucial role in regulating tumorangiogenesis (data not shown). The PGE2-induced angiogenesis(vWF positivity) was analyzed and represented in the form of abar graph (Fig. 5B). These data showed that PGE2-induced EP2and EP4 receptors mediated angiogenesis via NF-nB and Akt-dependent pathway and further suggested that both EP2 and EP4receptors might play important roles in regulating PGE2-inducedtumor angiogenesis.

Correlation between expression profiles of EP2 and EP4with ATF-4, c-Jun, c-Fos, uPA, and VEGF and their significancein prostate tumor progression. Our in vitro and in vivo dataprompted us to examine the expression profiles of variousoncogenic and angiogenic molecules in human prostate cancerspecimens by immunohistochemistry. The results showed that theexpression levels of ATF-4, c-Jun, c-Fos, uPA, and VEGF were higherin malignant tumors compared with normal and PIN specimens,which further correlated with the enhanced expressions of EP2 andEP4 in human prostate cancer specimens (Fig. 6A). The expressionsof these molecules were quantified and represented in the form of abar graph (Fig. 6A).

Discussion

Prostate cancer is considered as one of the most lethal diseasefor men in the United States and other parts of the world. To date,treatments like androgen deprivation therapy and chemotherapyare two of the major approaches known to increase survival ofpatients with metastatic prostate cancer; however, some sideeffects have been observed in patients undergoing these therapies(2, 35). Therefore, identification of novel prognostic marker anddevelopment of new therapeutic strategies could be the mostpromising approaches in the next generation of prostate cancermanagement.

Recently, several studies have shown the correlation betweenoverexpression of COX-2 with prostate tumorigenesis; however, themolecular mechanism underlying COX-2–induced prostate cancerprogression and angiogenesis is still not well understood. Althoughvarious reports have revealed the relationship between the elevatedlevels of PGE2 with malignant cancers (3), the molecularmechanism underlying PGE2-mediated prostate tumor progressionis still the subject of intense investigation. In this study, we haveshown the in-depth molecular mechanism underlying PGE2-induced tumor cell motility and angiogenesis in prostate cancervia EP2 and EP4 receptor–dependent pathway.

EGFR and MAPK-mediated activation of transcription factor AP-1 has been reported as one of the crucial signaling cascade thataffects tumor cell motility in various cancers, including themalignancies in prostate (36). Elevated expression of EGFR hasbeen observed in higher grades of prostate cancer, which furthercorrelate with poor clinical prognosis (37, 38). Earlier studiesreported that activation of EGFR in response to PGE2 leads to thephosphorylation of ERK, which, in turn, regulates downstreamsignaling events (36, 39). Moreover, it has also been observed thatPGE2 transactivates EGFR, which ultimately influences cellmigration, proliferation, and invasiveness in different cancermodels (36, 39, 40). In this paper, we have shown that PGE2induces the activation of EGFR and MAPK signaling cascade inprostate cancer cells. Recently, Wang and colleagues have shownthat among different PGE2 receptors, EP2 and EP4 predominantlyexpress PC-3 cells in androgen-independent prostate cancer (21).

Cancer Research


Here, we have shown that blocking of both PGE2 receptors (EP2and EP4) by their specific antagonists curbs the PGE2-mediatedactivation of EGFR and downstreams signaling cascades, whichultimately suppress the prostate tumor cell motility.

Integrin h3 has been shown to play critical roles in severaldistinct processes, such as tumor growth, metastasis, andangiogenesis in various cancers, including prostate cancer(41–43). Phosphorylation of h3 integrin is essential for theactivation of small GTP-binding proteins (Rho family), andactivation of Rho is necessary for invasion and migration in awide variety of cell types (44). Previous studies have indicated thatintegrins control activation of AP-1 in prostate cancer cells (26).Furthermore, earlier reports have shown that overexpression of h3integrin correlates with enhanced metastatic phenotype in LNCaPcells (42). Moreover, it has been observed that stromal cell derivedfactor-1 transiently increased the expression and activation of h3integrin in prostate cancer cells, which in turn augmented theaggressiveness of prostate cancer (43). In this study, we have

reported that PGE2 induces the expression and phosphorylation ofh3 integrin in PC-3 cells. The EP2 and EP4 receptor antagonistssuppressed PGE2-induced expression and phosphorylation of h3integrin, which further showed the involvement of both thesereceptors in this process. Silencing of h3 integrin expression by itsspecific siRNA suppresses PGE2-induced AP-1 activation, suggest-ing that PGE2 via EP2/EP4 controls AP-1 activation through h3integrin–mediated pathway. Chen and colleagues have shown thatarachidonic acid regulates PGE2-mediated PKA-dependent expres-sion of c-fos in PC-3 cells (5). In this study, we have shown themolecular mechanism, at least in part, whereby PGE2 via EGFR-ERK or h3 integrin–mediated pathway augments the expression ofc-Fos and phosphorylation of c-Jun, which ultimately regulatesAP-1 activation in prostate cancer cells.

Previous data showed that ATF-4 forms heterodimers with eitherc-Fos or c-Jun subunit of AP-1 and regulates the activation of AP-1(45). In this study, we have shown that PGE2 augments ATF-4activation, controls the interaction between ATF-4 and phosphor-

Figure 5. A, PGE2 regulates migration of endothelial cells toward the tumor cells. HUVEC, either alone or pretreated with EP2 or EP4 receptor antagonist, were used inthe upper chamber, whereas PC-3 cells, either alone or transfected with wt COX-2 cDNA or COX-2 siRNA, were used in the lower chamber. In separate experiments,PGE2 was used in the lower chamber as positive control. The endothelial cells migrated toward the reverse side of the upper chamber were stained with Giemsa,photographed, counted in 3 hpf, and represented in the form of a bar graph (*, P < 0.004; **, P < 0.015). B, PGE2 regulates EP2/EP4 receptor–dependent angiogenesisin Matrigel plug assay. PGE2 alone or along with AH6809 or AH23848 was mixed with Matrigel and injected s.c. in male athymic nude mice (NMRI, nu/nu). After 21 d,mice were sacrificed and Matrigel plugs were photographed and analyzed by immunohistochemistry using anti-CD31, anti-vWF, anti–phosphorylated p65, andanti–phosphorylated Akt antibodies. CD31 was stained with Cy2-conjugated IgG (green ), whereas vWF, phosphorylated p65, and phosphorylated Akt were stained withCy3-conjugated IgG (red). Nuclei were stained with DAPI (blue ). The vWF-positive staining was quantified by Image Pro Plus 6.0 Software and represented graphically.Columns, means of three determinations; bars, SE (*, P = 0.036).



ylated c-Jun, and regulates ATF-4–dependent AP-1 activation inprostate cancer cells.

uPA and its receptor uPAR-mediated signaling has beenimplicated in tumor cell invasion, survival, and metastasis in avariety of cancers including prostate (20). Both uPA and uPAR areexpressed at higher levels in malignant prostate tissues than inbenign and normal prostate tissues (20). VEGF acts as one of thekey proangiogenic factor responsible for neovascularization incancer cells (46). Previous studies have indicated that inhibition ofVEGF expression is a critical step in arresting prostate tumorgrowth and progression (47, 48). In this study, we have shownin-depth molecular mechanism underlying PGE2-induced EP2/EP4-mediated expression of uPA and VEGF via activation of AP-1 andATF-4, which eventually affects tumor cell motility and angiogen-esis. Moreover, our data revealed that inhibition of tumor-derivedPGE2 by COX-2 siRNA suppressed uPA and VEGF expression,suggesting that PGE2, both tumor-derived and exogenous, regulatesthis process in prostate cancer cells.

Angiogenesis is one of the most crucial steps in thedevelopment of tumor. The proliferation and migration ofendothelial cells play crucial roles in the regulation of tumor-associated angiogenesis (49). We have shown that PGE2 inducesthe phosphorylation of Akt and NF-nB, p65 in endothelialcells. Moreover, both exogenous and tumor-derived PGE2 aug-ments endothelial cell motility, and blocking of endothelial EP2and EP4 receptors by their antagonists suppresses endothelialcell motility toward tumor cells. In vivo Matrigel plugangiogenesis assay showed that PGE2 induces angiogenesiswhereas blocking EP2 and EP4 receptors suppressed this effect,indicating that PGE2 augments EP2/EP4-mediated in vivoangiogenesis. Our data also revealed that tumor-derived PGE2induces tumor angiogenesis. Taken together, these results showedthat PGE2 via EP2/EP4 receptor promotes in vivo angiogenesisthrough activation of Akt and NF-nB, suggesting that PGE2 playsan important role in the regulation of tumor angiogenesis.Furthermore, prostate tumor clinical specimen analysis data also

Figure 6. Expression profiles of EP2, EP4, ATF-4, c-Jun, c-Fos, uPA, and VEGF in human prostate cancer specimens and their correlation with human prostate cancerprogression in different pathologic grades. A, the levels of EP2, EP4, ATF-4, c-Jun, c-Fos, uPA, and VEGF were detected by immunohistochemical studies using theirspecific antibodies. EP2, EP4, ATF-4, c-Jun, c-Fos, and uPA were stained with Cy3-conjugated IgG (red ). VEGF was stained with Cy2-conjugated IgG (green ).Sections stained with anti-rabbit IgG were used as control. Nuclei were stained with DAPI (blue ). The expression profiles were quantified by Image Pro Plus 6.0Software and represented in the form of a bar graph (*, P < 0.003; **, P < 0.006; #, P < 0.02). B, schematic representation of PGE2-induced EP2/EP4-mediated EGFR/MAPK or h3 integrin–dependent ATF-4/AP-1 activation leading to enhanced uPA and VEGF expression, which in turn controls prostate tumor cell motility andangiogenesis. In endothelial cells, PGE2 through EP2/EP4 receptor stimulates Akt and NF-nB activation leading to enhanced endothelial cell motility and angiogenesis.

Cancer Research


corroborated with in vitro and in vivo findings, indicating higherlevels of uPA and VEGF expression in malignant prostate tumorscompared with normal and PIN tissues, which further correlatedwith the enhanced expression levels of ATF-4, c-Fos, c-Jun, EP2,and EP4. These data provide, at least in part, the molecular basisby which PGE2 controls downstream signaling cascades andleads to the expression of various oncogenic and angiogenicmolecules that ultimately regulate the prostate cancer progres-sion and angiogenesis (Fig. 6B). Thus, targeting PGE2 byinterfering its interaction with EP2 and EP4 receptors might bean alternative therapeutics that may aid in the rational design oftherapeutic strategy for the next generation of prostate cancertreatment.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

Received 12/17/2007; revised 6/18/2008; accepted 7/10/2008.Grant support: Council of Scientific and Industrial Research, Government of India

(S. Jain and G. Chakraborty).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 accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Ashwini Atre of National Center for Cell Science for the confocalmicroscopy–related experiments, Hemangini Shikhare of National Center for CellScience for flow cytometry–related work, and Dr. Tushar V. Patil, Department ofPathology, YCM Hospital, for clinical specimen–related work.



References

1. Pruthi RS, Derksen E, Gaston K. Cyclooxygenase-2 as apotential target in the prevention and treatment ofgenitourinary tumors: a review. J Urol 2003;169:2352–9.

2. Rostom A, Dube C, Lewin G, et al. Nonsteroidal anti-inflammatory drugs and cyclooxygenase-2 inhibitors forprimary prevention of colorectal cancer: a systematicreview prepared for the U.S. Preventive Services TaskForce. Ann Intern Med 2007;146:376–89.

3. Wallace JM. Nutritional and botanical modulation ofthe inflammatory cascade-eicosanoids, cyclooxygenases,and lipoxygenases-as an adjunct in cancer therapy.Integr Cancer Ther 2002;1:7–37.

4. Chaudry AA, Wahle KW, McClinton S, Moffat LE.Arachidonic acid metabolism in benign and malignantprostatic tissue in vitro : effects of fatty acids andcyclooxygenase inhibitors. Int J Cancer 1994;57:176–80.

5. Chen Y, Hughes-Fulford M. Prostaglandin E2 and theprotein kinase A pathway mediate arachidonic acidinduction of c-fos in human prostate cancer cells. Br JCancer 2000;82:2000–6.

6. Rao R, Redha R, Macias-Perez I, et al. ProstaglandinE2–4 receptor promotes endothelial cell migration viaERK activation and angiogenesis in vivo . J Biol Chem2007;282:16959–68.

7. Jain S, Chakraborty G, Kundu GC. The crucial role ofcyclooxygenase-2 in osteopontin-induced protein ki-nase C a/c-Src/InB kinase a/h-dependent prostatetumor progression and angiogenesis. Cancer Res 2006;66:6638–48.

8. Southall MD, Vasko MR. Prostaglandin receptorsubtypes, EP3C and EP4, mediate the prostaglandinE2-induced cAMP production and sensitization ofsensory neurons. J Biol Chem 2001;276:16083–91.

9. Liu X, Sun SQ, Ostrom RS. Fibrotic lung fibroblastsshow blunted inhibition by cAMP due to deficient cAMPresponse element-binding protein phosphorylation. JPharmacol Exp Ther 2005;315:678–87.

10. Montminy M. Transcriptional regulation by cyclicAMP. Annu Rev Biochem 1997;66:807–22.

11. Ameri K, Harris AL. Activating transcription factor 4.Int J Biochem Cell Biol 2008;40:14–21.

12. Rao JS. Molecular mechanisms of glioma invasive-ness: the role of proteases. Nat Rev Cancer 2003;3:489–501.

13. Miyake H, Hara I, Yamanaka K, Arakawa S, KamidonoS. Elevation of urokinase-type plasminogen activatorand its receptor densities as new predictor of diseaseprogression and prognosis in men with prostate cancer.Int J Oncol 1999;14:535–41.

14. Li Y, Cozzi PJ. Targeting uPA/uPAR in prostatecancer. Cancer Treat Rev 2007;33:521–7.

15. Delongchamps NB, Peyromaure M, Dinh-Xuan AT.Role of vascular endothelial growth factor in prostatecancer. Urology 2006;68:244–8.

16. McCarty MF. Targeting multiple signaling pathwaysas a strategy for managing prostate cancer; multifocalsignal modulation therapy. Integr Cancer Ther 2004;3:349–80.

17. Garrido C, Saule S, Gospodarowicz D. Transcriptionalregulation of vascular endothelial growth factor gene

expression in ovarian bovine granulosa cells. GrowthFactors 1993;8:109–17.

18. Jeon SH, Chae BC, Kim HA, et al. The PKA/CREBpathway is closely involved in VEGF expression inmouse macrophages. Mol Cells 2007;23:23–9.

19. Mira-Y-Lopez R, Jaramillo S, Jing Y. Synergistictranscriptional activation of the mouse urokinaseplasminogen activator (uPA) gene and of its enhanceractivator protein 1 (AP1) site by cAMP and retinoic acid.Biochem J 1998;331:909–16.

20. Gavrilov D, Kenzior O, Evans M, Calaluce R, Folk WR.Expression of urokinase plasminogen activator andreceptor in conjunction with the ets family and AP-1complex transcription factors in high grade prostatecancers. Eur J Cancer 2001;37:1033–40.

21. Wang X, Klein RD. Prostaglandin E2 induces vascularendothelial growth factor secretion in prostate cancercells through EP2 receptor-mediated cAMP pathway.Mol Carcinog 2007;46:912–23.

22. Chakraborty G, Jain S, Kundu GC. Osteopontinpromotes vascular endothelial growth factor-dependentbreast tumor growth and angiogenesis via autocrine andparacrine mechanisms. Cancer Res 2008;66:152–61.

23. Chakraborty G, Rangaswami H, Jain S, Kundu GC.Hypoxia regulates cross-talk between Syk and Lckleading to breast cancer progression and angiogenesis.J Biol Chem 2006;281:11322–31.

24. Tang CH, Yang RS, Fu WM. Prostaglandin E2stimulates fibronectin expression through EP1 receptor,phospholipase C, protein kinase Ca, and c-Src pathwayin primary cultured rat osteoblasts. J Biol Chem 2005;280:22907–16.

25. Mayoral R, Fernandez-Martinez A, Bosca L, Martin-Sanz P. Prostaglandin E2 promotes migration andadhesion in hepatocellular carcinoma cells. Carcinogen-esis 2005;26:753–61.

26. Quyang X, Jessen WJ, Al-Ahmadie H, et al. Activatorprotein-1 transcription factors are associated withprogression and recurrence of prostate cancer. CancerRes 2008;68:2132–44.

27. Kaur M, Agarwal R. Transcription factors: moleculartargets for prostate cancer intervention by phytochem-icals. Curr Cancer Drug Targets 2007;7:355–67.

28. Wisdom R. AP-1: one switch for many signals. ExpCell Res 1999;253:180–5.

29. Karin M. The regulation of AP-1 activity by mitogen-activated protein kinases. J Biol Chem 1995;270:16483–6.

30. Gupta S, Barrett T, Whitmarsh AJ, et al. Selectiveinteraction of JNK protein kinase isoforms withtranscription factors. EMBO J 1996;15:2760–70.

31. Attiga FA, Fernandez PM, Weeraratna AT, ManyakMJ, Patierno SR. Inhibitors of prostaglandin synthesisinhibit human prostate tumor cell invasiveness andreduce the release of matrix metalloproteinases. CancerRes 2000;60:4629–37.

32. Albini A, Dell’Eva R, Vene R, et al. Mechanisms ofthe antiangiogenic activity by the hop flavonoidxanthohumol: NF-nB and Akt as targets. FASEB J 2006;20:527–9.

33. Shiojima I, Walsh K. Role of Akt signaling invascular homeostasis and angiogenesis. Circ Res 2002;90:1243–50.

34. Masferrer JL, Koki A, Seibert K. COX-2 inhibitors:anew class of antiangiogenic agents. Ann N Y Acad Sci1999;889:84–6.

35. Higano C. Androgen deprivation therapy: monitoringand managing the complications. Hematol Oncol ClinNorth Am 2006;20:909–23.

36. Battersby S, Sales KJ, Williams AR, Anderson RA,Gardner S, Jabbour HN. Seminal plasma and prosta-glandin E2 up-regulate fibroblast growth factor 2expression in endometrial adenocarcinoma cells via E-series prostanoid-2 receptor-mediated transactivationof the epidermal growth factor receptor and extracel-lular signal-regulated kinase pathway. Hum Reprod2007;22:36–44.

37. Prewett M, Rockwell P, Rockwell RF, et al. Thebiologic effects of C225, a chimeric monoclonalantibody to the EGFR, on human prostate carcinoma.J Immunother Emphasis Tumor Immunol 1996;19:419–27.

38. Hernes E, Fossa SD, Berner A, Otnes B, Nesland JM.Expression of the epidermal growth factor receptorfamily in prostate carcinoma before and duringandrogen-independence. Br J Cancer 2004;90:449–54.

39. Buchanan FG, Wang D, Bargiacchi F, DuBois RN.Prostaglandin E2 regulates cell migration via theintracellular activation of the epidermal growth factorreceptor. J Biol Chem 2003;278:35451–7.

40. Thomas SM, Bhola NE, Zhang Q. Cross-talk betweenG protein-coupled receptor and epidermal growthfactor receptor signaling pathways contributes togrowth and invasion of head and neck squamous cellcarcinoma. Cancer Res 2006;66:11831–9.

41. Cooper CR, Chay CH, Pienta KJ. The role of a(v)h(3)in prostate cancer progression. Neoplasia 2002;4:191–4.

42. Zheng DQ, Woodard AS, Fornaro M, Tallini G,Languino LR. Prostatic carcinoma cell migration viaa(v)h3 integrin is modulated by a focal adhesion kinasepathway. Cancer Res 1999;59:1655–64.

43. Sun YX, Fang M, Wang J, Cooper CR, Pienta KJ,Taichman RS. Expression and activation of avh3integrins by SDF-1/CXC12 increases the aggressivenessof prostate cancer cells. Prostate 2007;67:61–73.

44. Butler B, Williams MP, Blystone SD. Ligand-depen-dent activation of integrin avh3. J Biol Chem 2003;278:5264–70.

45. Hai T, Curran T. Cross-family dimerization oftranscription factors Fos/Jun and ATF/CREB altersDNA binding specificity. Proc Natl Acad Sci U S A1991;88:3720–4.

46. McMahon G VEGF receptor signaling in tumorangiogenesis. Oncologist 2000;5:3–10.

47. Bocci G, Man S, Green SK, et al. Increased plasmavascular endothelial growth factor (VEGF) as a surro-gate marker for optimal therapeutic dosing of VEGFreceptor-2 monoclonal antibodies. Cancer Res 2004;64:6616–25.

48. Byrne AM, Bouchier-Hayes DJ, Harmey JH. Angio-genic and cell survival functions of vascular endothelialgrowth factor (VEGF). J Cell Mol Med 2005;9:777–94.

49. Isner JM, Asahara T. Angiogenesis and vasculo-genesis as therapeutic strategies for postnatal neo-vascularization. J Clin Invest 1999;103:1231–6.

Hypoxia Regulates Cross-talk between Syk and Lck Leadingto Breast Cancer Progression and Angiogenesis*

Received for publication, November 23, 2005, and in revised form, February 3, 2006 Published, JBC Papers in Press, February 10, 2006, DOI 10.1074/jbc.M512546200

Goutam Chakraborty1, Hema Rangaswami1, Shalini Jain, and Gopal C. Kundu2

From the National Center for Cell Science, Pune 411 007, India

Hypoxia is a key parameter that controls tumor angiogenesis andmalignant progression by regulating the expression of several onco-genicmolecules. The nonreceptor protein-tyrosine kinases Syk andLck play crucial roles in the signalingmechanism of various cellularprocesses. The enhanced expression of Syk in normal breast tissuebut not in malignant breast carcinoma has prompted us to investi-gate its potential role in mammary carcinogenesis. Accordingly, wehypothesized that hypoxia/reoxygenation (H/R)mayplay an impor-tant role in regulating Syk activation, and Lck may be involved inthis process. In this study, we have demonstrated that H/R differen-tially regulates Syk phosphorylation and its subsequent interactionand cross-talk with Lck inMCF-7 cells. Moreover, Syk and Lck playdifferential roles in regulating Sp1 activation and expressions ofmelanoma cell adhesion molecule (MelCAM), urokinase-type plas-minogen activator (uPA), matrix metalloproteinase-9 (MMP-9),and vascular endothelial growth factor (VEGF) in response to H/R.Overexpression of wild type Syk inhibited the H/R-induced uPA,MMP-9, and VEGF expression but up-regulated MelCAM expres-sion. Our data also indicated that MelCAM acts as a tumor sup-pressor by negatively regulating H/R-induced uPA secretion andMMP-9 activation. Themice xenograft study showed the cross-talkbetween Syk and Lck regulated H/R-induced breast tumor progres-sion and further correlated with the expressions of MelCAM, uPA,MMP-9, and VEGF. Human clinical specimen analysis supportedthe in vitro and in vivo findings. To our knowledge, this is firstreport that the cross-talk between Syk and Lck regulates H/R-in-duced breast cancer progression and further suggests that Syk mayact as potential therapeutic target for the treatment of breastcancer.

Hypoxia plays a crucial role in regulating breast tumor progressionthrough a multistep process that includes oncogene activation or inhi-bition of tumor suppressor genes (1). Most tumors develop regions ofchronically or transiently hypoxic cells during growth (2). Hypoxictumor regionsmay show increased expression ofmany genes because ofhypoxia-induced activation of transcription factors (3–5). Low extracel-lular pH, glucose depletion, high lactate levels, and regions with lowoxygen tension (6, 7) characterize most tumors. Low oxygen tension intumors has been associated with poor outcome, enhanced local orlocoregional spread, and enhancedmetastatic potential (8). Hypoxia is akey parameter, whichmodulates the expression of a variety of genes thatare involved in tumor angiogenesis, malignant progression, and distant

metastasis (9). The signaling properties of reactive oxygen species arebecause of the reversible oxidation of redox-sensitive target proteins(10). The generated reactive oxygen species act as intracellular secondmessengers in various signal transduction pathways and hence play acrucial role in regulating disease and stress-induced cellular injuriessuch as ischemia/reperfusion, UV irradiation, and inflammation (10).Previous reports have indicated that areas of hypoxia/reoxygenation

(H/R)3 are a typical feature of rapidly growing andmetastasizing tumors(11, 12). It was also demonstrated that both hypoxia and consecutivehypoxia/reoxygenation exert a variety of influence in tumor cell biologythat ultimately regulates tumor progression (13). Earlier reports alsoshowed that hypoxia and H/R regulate the activation of various mito-gen-activated protein kinase signaling pathways (13) and induce theactivation of several transcriptional factors such as HIF-1�, NF�B,AP-1, and Sp1 (13–15).However, very recently it was demonstrated thatH/R rather than hypoxia alone appears to induce the expression andactivation of several oncogenicmolecules and plays an important role intumor progression (13, 16).The nonreceptor protein-tyrosine kinase Syk is widely expressed in

hematopoietic cells (17, 18). It has tandem amino-terminal Src homol-ogy 2 domains and a carboxyl-terminal kinase domain (19). The Srchomology 2 domains bind phosphorylated immunoreceptor tyrosine-based activation motifs and hence play a significant role in immunore-ceptor and cytokine signaling (20). The expression of Syk has also beenreported in cell lines of epithelial origin (21), but its function in thesecells is not well understood. It has been documented that Syk is com-monly expressed in normal humanbreast tissue, benign breast lesions, andlow tumorigenic breast cancer cell lines (22). Previous data indicated thatSyk suppresses cell motility and NF�B-mediated urokinase-type plas-minogen activator (uPA) secretion by inhibiting phosphatidylinositol3-kinase activity in breast cancer cells (23). Lck, a member of the Srcfamily nonreceptor protein-tyrosine kinase, is mostly expressed in Tcells, breast cancer tissues, and cell lines and also in some B cells (24).Lck binds to the cytoplasmic domain of CD4 and CD8 and plays anessential role in T cell activation and development (25). Earlier reportshave indicated that p72Syk plays a crucial role in activation of p56Lck

through physical association and amino-terminal tyrosine phosphoryl-ation at residues Tyr-518 and Tyr-519. Mutation of these residues tophenylalanines abolished its activity in vitro and toward cellular sub-strates in vivo and reduced its tyrosine phosphorylation by �90% (26).However, the molecular mechanism by which H/R regulates Syk phos-phorylation and its subsequent interaction with Lck leading to down-stream signaling events in breast cancer cells are not well defined.

* The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked “advertisement” in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

1 Both authors contributed equally to this work.2 To whom correspondence should be addressed: National Center for Cell Science, NCCS

Complex, Pune 411 007, India. Tel.: 91-20-25690922 (ext. 1103); Fax: 91-20-25692259;E-mail: [email protected].

3 The abbreviations used are: H/R, hypoxia/reoxygenation; Syk, splenic tyrosine kinase;Lck, leukocyte-specific kinase; HIF-1�, hypoxia-inducible factor1�; EMSA, electro-phoretic mobility shift assay; MelCAM, melanoma cell adhesion molecule; uPA, uroki-nase-type plasminogen activator; MMP-9, matrix metalloproteinase-9; VEGF, vascularendothelial growth factor; vWF, von Willebrand factor; FITC, fluorescein isothiocya-nate; TRITC, tetramethylrhodamine isothiocyanate; DAPI, 4,6-diamidino-2-phenylin-dole; WT, wild type; Mut, mutant; DN, dominant negative; NMRI, Naval MedicalResearch Institute.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 281, NO. 16, pp. 11322–11331, April 21, 2006© 2006 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

11322 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 16 • APRIL 21, 2006

MelCAM, previously known as MUC18 or MCAM, a newly recog-nized cell adhesion molecule with an apparent molecular mass of 113kDa, belongs to the immunoglobulin superfamily (27–29). The pres-ence of putative binding sites for the transcription factors Sp1, AP-1,AP-2, and cAMP-response element-binding protein in the promoterregion suggests that MelCAM expression can be modulated by exoge-nous factors (30). At the cellular level, phorbol ester and cyclic AMPhave been shown tomodulateMelCAM expression (31). Earlier reportshave indicated that MelCAM acts differently in the progression ofbreast carcinomas. It is expressed in normal and benign proliferativebreast epithelium, and its expression is frequently lost in in situ andinfiltrating breast carcinomas (32, 33). The MelCAM core promotercontains four binding sites for the Sp1 transcription factor, and deletionanalyses have indicated that removal of all putative Sp1 sites reduced thepromoter activity by 80%, suggesting that Sp1 is an important regulatorof MelCAM expression (34). However, the molecular mechanism bywhich H/R regulates Syk activation and Syk-dependent Lck-mediatedSp1 activation leading to the regulation of MelCAM expression inMCF-7 cells is not well defined.Degradation of extracellular matrix plays an important role in tumor

metastasis. uPA is a member of the serine protease family that interactswith uPA receptor and facilitates the conversion of inert zymogen plas-minogen into widely acting serine protease plasmin (35, 36). MMP-9,also referred as gelatinase-B, is not only associated with invasion andmetastasis but also has been implicated in angiogenesis, rheumatoidarthritis, retinopathy, and vascular stenosis and hence is considered tobe a prioritized therapeutic target (37, 38). Several reports have indi-cated the positive correlation between uPA/MMP-9 activation and themetastatic potential of tumors (39). However, the molecular mecha-nism by which H/R regulates Syk/Lck-dependent MelCAM expressionand uPA secretion and the uPA-dependent pro-MMP-9 activation inbreast carcinoma cells is not well understood. Moreover, the roles ofthese molecules in regulating H/R-induced tumorigenesis and its clini-cal implications are not well defined.Hypoxia-induced VEGF production provides one of the main

driving forces that stimulate the angiogenesis, which accompaniestumor progression (40, 41). To date, VEGF is considered as the keyfactor that guided and regulated tumor angiogenesis (42). Tumorcell-derived VEGF binds to its specific receptors and regulates tumorprogression through neovascularization via autocrine and paracrinepathways (43, 44). Recent evidence suggested that the VEGF pro-moter contains an Sp1-response element (45). However, the role ofLck and Syk in H/R-induced VEGF expression in breast cancer is notdefined clearly.In this study we have demonstrated the differential role of Syk and

Lck in the H/R-induced uPA, MMP-9, and VEGF expression. Ourfindings suggested that H/R down-regulates Syk activation leading toenhanced uPA, MMP-9, and VEGF expression. Furthermore, over-expression of Syk restored H/R-induced down-regulation of Mel-CAM expression. In hypoxic cells, Lck also physically associates withSyk, and this association plays a crucial role in regulating the down-stream signaling. The in vivo relevance of our study was furthervalidated in a xenografted nude mice model, which also supports ourin vitro findings. Clinical data also indicated that the higher grades oftumors showed significant HIF-1� expression compared with that oflower grades or normal breast tissue and also demonstrated aninverse correlation between Syk/MelCAM and uPA/MMP-9/VEGFexpression, which further correlates with enhanced tumorigenicpotential and neovascularization.

EXPERIMENTAL PROCEDURES

Materials—The rabbit polyclonal anti-Syk, anti-Lck, anti-Sp1, anti-MelCAM, anti-uPA, anti-MMP-9, anti-actin, mouse monoclonal anti-Lck, anti-Syk, anti-phosphotyrosine antibodies, andMelCAM blockingpeptide were purchased from Santa Cruz Biotechnology. Rabbit poly-clonal anti-VEGF was from Oncogene. Rabbit polyclonal anti-vWFantibody was from Sigma. Rabbit polyclonal anti-HIF-1� was fromUpstate Biotechnology, Inc. Lipofectamine Plus was obtained fromInvitrogen. pp2, aminogenistein, and damnacanthal were from Calbio-chem. The Sp1 consensus oligonucleotide was purchased from Banga-lore Genei. Matrigel was purchased from BD Biosciences. The[�-32P]ATP was purchased from the Board of Radiation and IsotopeTechnology (Hyderabad, India). The female nude mice (NMRI, nu/nu)were from National Institute of Virology (Pune, India). All other chem-icals were of analytical grade.

Cell Culture—TheMCF-7 cells were cultured inDulbecco’smodifiedEagle’s medium supplemented with 10% fetal calf serum, 100 units/mlpenicillin, 100 mg/ml streptomycin, and 2 mM glutamine in a humidi-fied atmosphere of 5% CO2 and 95% air at 37 °C.

Hypoxic/Reoxygenation Cultures—The MCF-7 cells grown to 50–70%confluence were made hypoxic in evacuation chambers by intermittentapplication of vacuum and sparging with 95% N2, 5% CO2. Cells wereanalyzed at this point or maintained under hypoxic conditions in thepresence of 100 mM dithionate (an O2 scavenger) at 37 °C for the indi-cated time point. These cells were reoxygenated for the indicated peri-ods by replacing the medium with fresh medium and incubating thecultures in humidified atmosphere of 5% CO2 and 95% air at 37 °C.

Immunofluorescence and Immunohistochemistry—To detect theeffect of H/R on cellular localization of Syk, cells grown inmonolayer onglass slides were induced by hypoxia for 2 h and reoxygenated for 45min. The cells were fixed and incubated with rabbit polyclonal anti-Sykantibody (1:50 dilution) followed by FITC-conjugated anti-rabbit IgG atroom temperature. The role of H/R in regulating Syk-Lck colocalizationwas determined by immunofluorescence studies using a mixture ofmouse monoclonal anti-Syk and rabbit polyclonal anti-Lck antibodyfollowed by a mixture of TRITC- and FITC-conjugated IgG. The cellswere washed and mounted with coverslips. All these samples were ana-lyzed under confocal microscopy (Zeiss).The clinical specimens were analyzed by immunohistopathological

studies. Formalin-fixed paraffin-embedded sections (4 �m) were sub-jected to antigen retrieval, and the Syk-Lck colocalization was deter-mined by immunofluorescence studies using a mixture of rabbit poly-clonal anti-Syk and mouse monoclonal anti-Lck antibodies followed bya mixture of FITC- and TRITC-conjugated IgGs. The levels of HIF-1�

expression in these clinical samples were determined by Western blotanalysis. The levels of MelCAM, uPA, MMP-9, and VEGF expressionsin these sampleswere detected by immunofluorescence studies by usingtheir specific antibodies. The tumormicrovessel densitieswere detectedby immunostaining with anti-vWF antibody. All these samples wereanalyzed under confocal microscopy (Zeiss).

Plasmids and DNA Transfection—The dominant-negative form ofLck (DN Lck, K273R) in pcDNA3 was a kind gift from Dr. D. R. Branch(Canadian Blood Services, Toronto, Ontario). Thewild type and kinase-negative Syk cDNA in pcDNA 3.1 were the generous gifts from Dr.Susette C. Mueller (Department of Oncology, Georgetown UniversityMedical School,Washington, D. C.). MCF-7 cells were transfected withspecific cDNAusing Lipofectamine Plus reagent (Invitrogen) accordingto the manufacturer’s instructions as described previously (46). Thesetransfected cells were used for Syk-Lck colocalization andinteraction studies, detection of Sp1, Syk, Lck, MelCAM, uPA,

Cross-talk between Syk and Lck in Response to H/R

APRIL 21, 2006 • VOLUME 281 • NUMBER 16 JOURNAL OF BIOLOGICAL CHEMISTRY 11323

MMP-9, andVEGF byWestern blot analysis, and in vivo tumorigenicityexperiments.

Zymography Experiments—The gelatinolytic activity was measuredas described (46, 47). To examine whether H/R regulates MMP-9 acti-vation and to investigate the effect of inhibition ofMelCAM expressionon H/R-inducedMMP-9 activation, cells were either induced with H/Rfor 24 h or pretreated withMelCAM blocking peptide (50 �g) and thenexposed with H/R. The conditionedmedia were collected, and the sam-ples were analyzed by zymography as described (47). Negative stainingshowed the zones of gelatinolytic activity.

Immunoprecipitation—To delineate the role of H/R in the regulationof tyrosine phosphorylation of Syk, cells were exposed to hypoxia for 2 hfollowed by reoxygenation for 0–60min. In separate experiments, cellswere pretreated with Lck inhibitors, pp2 (4 nM), aminogenistein (2 �M),and damnacanthal (0.8 �M), and then exposed to H/R. Cell lysates wereimmunoprecipitated with rabbit polyclonal anti-Syk antibody and ana-lyzed by Western blot using anti-phosphotyrosine antibody. The sameblots were reprobed with anti-Syk antibody. To analyze whether Sykinteracts with Lck and to determinewhetherH/R regulates this process,cells were exposed to H/R for 45 min. In other experiments, cells weretransfected with wild type and kinase-negative Syk and then subjectedto H/R. Cells were lysed in lysis buffer (50 mM Tris-HCl (pH 7.4), 150mMNaCl, 1%Nonidet P-40, 1%TritonX-100, 1% sodium deoxycholate,0.1% SDS, 5mM iodoacetamide, 2mMphenylmethylsulfonyl fluoride, 10�g/ml aprotinin, 1 �g/ml leupeptin, 1 �g/ml pepstatin). The lysatescontaining equal amount of total proteins were immunoprecipitatedwith rabbit anti-Lck antibody and analyzed byWestern blot using anti-Syk antibody. The same blots were reprobed with anti-Lck antibody.

Western Blot Analysis—To analyze the roles of Syk and Lck in regu-lating MelCAM, VEGF expression, and MMP-9 activation, cells weretransfected with wild type and kinase-negative Syk or dominant-nega-tive Lck and then exposed to H/R for 24 h as described earlier. The celllysates were analyzed by Western blot using anti-MelCAM or anti-VEGF antibody. The levels of Lck and Syk in nontransfected or trans-fected cell lysates were also detected by Western blot using anti-Lck oranti-Syk antibody. The level of activeMMP-9 in conditionedmedia wasdetected by Western blot.To analyze the effect of inhibition of MelCAM on H/R-induced uPA

secretion, cells were pretreated with MelCAM-specific blocking pep-tide (50 �g/ml) for 24 h and subjected to H/R treatment. Cell lysateswere analyzed by Western blot using anti-uPA antibody. The level ofHIF-1� expression in the normal and breast tumor specimens of differ-ent grades was also detected by Western blot using anti-HIF-1� anti-body. The same blots were reprobed with anti-actin antibody as loadingcontrol.

Nuclear Extracts andWestern Blot—TocheckwhetherH/R regulatesSp1 expression, cells were subjected toH/R for 0–180min at 37 °C. Thenuclear extracts were prepared as described earlier (24). Briefly, cellswere incubated in hypotonic buffer (10 mM HEPES (pH 7.9), 1.5 mM

MgCl2, 10 mM KCl, 0.2 mM phenylmethylsulfonyl fluoride, and 0.5 mM

dithiothreitol) and allowed to swell on ice for 10min. Cells were homog-enized in a Dounce homogenizer. The nuclei were separated by spin-ning at 3300 � g for 5 min at 4 °C. The supernatant was used as cyto-plasmic extract. The nuclear pellet was extracted in nuclear extractionbuffer (20 mM HEPES (pH 7.9), 0.4 M NaCl, 1.5 mM MgCl2, 0.2 mM

EDTA, 25% glycerol, 0.5mMphenylmethylsulfonyl fluoride, and 0.5mM

dithiothreitol) and centrifuged at 12,000 � g for 30 min. The superna-tant was used as nuclear extract. The nuclear extracts were resolved bySDS-PAGE, and the level of Sp1 was detected by Western blot usingrabbit anti-Sp1 antibody. In separate experiments, to examine the

effects of Syk and Lck on H/R-induced Sp1 expression, cells were trans-fected with wild type and kinase-negative Syk or DN Lck and theninduced with H/R. The nuclear extracts were prepared and analyzed byWestern blot using anti-Sp1 antibody. The levels of Lck and Syk innuclear extracts were also detected by Western blot using anti-Lck oranti-Syk antibody. The level of expression of HIF-1� in hypoxia or H/R-induced nuclear extracts was detected by Western blot using anti-HIF-1� antibody. Actin was used as loading control.

EMSA—To examine whether H/R induces the Sp1-DNA binding, cellswere exposed toH/R for 0–180min as described earlier. To checkwhetherSyk and Lck play any role in regulating H/R-induced Sp1-DNA binding,cells were transfected with wild type and kinase-negative Syk or pretreatedwith Lck inhibitors, pp2 (4 nM), aminogenistein (2�M), and damnacanthal(0.8 �M) followed by exposed to H/R for 1 h. The nuclear extracts wereprepared as described above and incubated with 16 fmol of 32P-labeleddouble-stranded Sp1 oligonucleotide (5�-ATTCGATCGGGGCGGGGC-3�) in binding buffer (25mMHEPES (pH 7.9), 0.5mMEDTA, 0.5mMdithi-othreitol, 1% Nonidet P-40, 5% glycerol, 50 mM NaCl) containing 1 �g ofpoly(dI-dC). The DNA-protein complex was resolved on a native poly-acrylamide gel and analyzed by autoradiography.

InVivoXenograftTumorModel—The tumorigenicity experimentswereperformed as described (47, 48). Briefly, female athymic nudemice, NMRI,nu/nu (BALB/c) obtained fromNational Institute ofVirology (Pune, India)were housed under specific pathogen-free conditions and used for in vivotumorigenicity studies. Cells were either exposed to H/R or transfectedwithwild typeormutantSykand thenexposed toH/R for24h.Cell viabilitywas determined by the trypan blue exclusion test, and only a single cellsuspensions (5 � 106/0.2 ml) of �90% viability were mixed with Matrigeland injected subcutaneously into the flanks of female athymic NMRI (nu/nu)mice (6–8 weeks old). Fivemice were used in each set of experiments.Growth of tumors was monitored weekly by measuring the tumors withcalipers. The mice were killed after 4 weeks of injection, and the levels ofuPA, VEGF, and MelCAM in the tumors were analyzed by Western blot.The level of Sp1-DNAbinding in the nuclear extracts of the tumor sampleswas determined by EMSA.The tumor samples were also processed for histopathological studies.

Formalin-fixed paraffin-embedded sections (4 �m) were subjected toantigen retrieval and then stained with rabbit polyclonal anti-MMP-9,anti-uPA antibody, or mouse monoclonal anti-MelCAM antibody fol-lowed by FITC-conjugated anti-rabbit or anti-mouse IgG. The tumormicrovessel densities were detected by immunostaining with anti-vWFantibody. The level of colocalization of Syk and Lck in these tumorspecimens was determined by immunofluorescence using a mixture ofmouse monoclonal anti-Syk and rabbit polyclonal anti-Lck antibodyfollowed by a mixture of FITC- and TRITC-conjugated IgG. All thesesamples were analyzed under confocal microscopy (Zeiss).

Human Breast Tumor Specimen Analysis—Human breast tumor speci-menswith different grades and normal breast tissueswere collected from alocal hospital with informed consent and were flash-frozen. The immuno-histochemical and Western blot analyses of those specimens were per-formed as described above.

RESULTS

H/R Regulates the Cellular Localization of Syk—To determinewhether H/R regulates the cellular localization of Syk,MCF-7 cells (2�105 cells/ml) seeded on 35-mm tissue culture plate were exposed tohypoxia for 2 h and then reoxygenated for 45 min. After that, cells werefixed and incubated with rabbit polyclonal anti-Syk antibody followedby FITC-conjugated anti-rabbit IgG. The data indicated that H/Rinduces the cytoplasmic translocation of Syk at 45 min in these cells



(Fig. 1A). To further confirm the cellular localization of Syk in responseto H/R, cells were exposed to hypoxia for 2 h followed by reoxygenationfor 0–60min. The cytoplasmic and nuclear fractions were isolated, andthe level of Syk was detected byWestern blot. The results indicated thatcytoplasmic translocation of Sykwas found at 45min in response toH/Rexposure (Fig. 1B). The same blots were reprobed with anti-actin anti-body as loading control. Previous studies demonstrated that HIF-1� isconsidered as key transcription factor and acts asmarker under hypoxicconditions (15). Accordingly cells were exposed to hypoxia for 0–4 h orexposed to hypoxia for 2 h and then reoxygenated for 0–60 min. Thenuclear extracts were prepared and subjected to Western blot analysisusing anti-HIF-1� antibody. The results indicated that nuclear expres-sion of HIF-1� was found after 4 h of exposure with hypoxia (Fig. 1C,lanes 1–5), whereas maximum nuclear expression of HIF-1� wasdetected when cells were exposed to hypoxia for 2 h and reoxygenationfor 30 min and rapidly returned to basal level at 60 min (Fig. 1D, lanes1–5). The level of expression ofHIF-1� in response to hypoxia or reoxy-genation in nuclear extracts was detected as positive control. Actin wasused as loading control.

H/R Suppresses Syk Phosphorylation—Toascertain the role ofH/R ontyrosine phosphorylation of Syk, cells were exposed to hypoxia for 2 hand then reoxygenated for 0–60 min. The cell lysates were immuno-precipitated with rabbit polyclonal anti-Syk antibody and analyzed byWestern blot using anti-phosphotyrosine antibody (Fig. 2A, upperpanel, lanes 1–5). The data revealed that cells exposed to H/R sup-pressed tyrosine phosphorylation of Syk at 45min (Fig. 2A, upper panel,lane 4) and were restored at 60 min (lane 5). These data suggested thatdephosphorylation of Syk at 45 min leads to cytoplasmic translocation,whereas rephosphorylation causes its nuclear translocation at 60 min(Figs. 1B and 2A). The same blots were reprobed with anti-Syk antibody(Fig. 2A, lower panel).To delineate whether p56Lck plays any role in Syk phosphorylation in

response to H/R, cells were pretreated with Lck inhibitors (pp2, amin-ogenistein, and damnacanthal) and then treated with H/R. Cell lysateswere immunoprecipitated with anti-Syk antibody and immunoblottedwith anti-Tyr(P) antibody (Fig. 2B, upper panel, lanes 1–5). The sameblots were reprobed with anti-Syk antibody (Fig. 2B, lower panel). The

results indicated that cells treated with Lck inhibitors restored the Sykphosphorylation, which was down-regulated upon exposure to H/R.

H/R Regulates the Interaction between Syk and Lck—To examine therole of H/R in regulating the interaction between Syk and Lck, cells wereeither exposed to H/R or transfected with wild type andmutant Syk andthen exposed to H/R. Cell lysates were immunoprecipitated with anti-Lck antibody and analyzed byWestern blot with anti-Syk antibody (Fig.2C, upper panel). Cells transfected with WT Syk but not with Mut Sykresulted in restoration of Syk and Lck interaction in response to H/R.The same blots were reprobed with anti-Lck antibody (Fig. 2C, lowerpanel). The cellular colocalization between Syk and Lck in response toH/R was further confirmed by immunofluorescence study (Fig. 2D).These results suggested that H/R down-regulates the physical associa-tion between Syk and Lck, and this is due to cytoplasmic translocation ofSyk in response to H/R at 45 min.

SykNegatively RegulatesH/R-induced Sp1Activation—Todeterminethe role of H/R on Sp1 nuclear translocation and DNA binding, MCF-7cells were exposed to H/R for 0–180 min as described earlier, and thecytoplasmic and nuclear extracts were prepared. Nuclear extracts weresubjected toWestern blot using anti-Sp1 antibody (Fig. 3A, upper panel,lanes 1–5). Similarly, cytoplasmic extracts were also used for Westernblot using anti-Sp1 antibody (Fig. 3A, 2ndmiddle panel). Actinwas usedas loading control (Fig. 3A, 1st middle and lower panels). To examinewhether H/R regulates Sp1-DNA binding, nuclear extracts were ana-lyzed by EMSA using Sp1-specific oligonucleotide (Fig. 3C, lanes 1–5).The results showed that H/R induces maximum Sp1 nuclear transloca-tion and DNA binding at 60 min (Fig. 3, A and C). To examine the rolesof Syk and Lck inH/R-induced Sp1 nuclear translocation and Sp1-DNAbinding, cells were individually transfected with WT and Mut Syk orDN Lck or treated with Lck-specific inhibitors (pp2, aminogenistein,and damnacanthal) and were then exposed to H/R. Nuclear extractswere analyzed by Western blot (Fig. 3B, upper panel, lanes 1–5) andEMSA (Fig. 3, D and E). The blots were reprobed with anti-Syk oranti-Lck antibody in order to detect the expression of Syk and Lck (Fig.3B, 1st middle and 2nd middle panels). Actin was used as loading con-trol (Fig. 3B, lower panel). The results demonstrated that Syk negativelyregulates H/R-induced Sp1 nuclear translocation and DNA binding,

FIGURE 1. H/R regulates cellular localization ofSyk. A, MCF-7 cells were exposed to hypoxia for 2 hand reoxygenated for 45 min. The cellular localiza-tion of Syk was determined by immunofluorescencestaining using anti-Syk antibody and analyzed byconfocal microscopy. a and b, imunofluorescencephotographs; a’ and b’, phase contrast photographs.B, cellular localization of Syk was analyzed by West-ern blot. Cells were exposed to hypoxia for 2 h andthen reoxygenated for 0–60 min. Cytoplasmic andnuclear extracts were prepared, and the cellularlocalization of Syk was detected by Western blotusing anti-Syk antibody. Actin was used as loadingcontrol. Note that maximum level of cytoplasmictranslocation of Syk was observed at 45 min of expo-sure with H/R, and its least expression was noticed inthe nucleus at 45 min. C and D, effects of hypoxia andH/R on nuclear localization of HIF-1�. Cells wereexposed to hypoxia for 0–4 h or exposed to hypoxiafor 2 h followed by reoxygenation for 0–60 min. Thelevel of HIF-1� in the nuclear extracts was detectedby Western blot. Note that the higher level of HIF-1�was observed at 4 h of hypoxic treatment or at 30min of H/R exposure. Actin was used as loading con-trol. The results shown here represent three experi-ments exhibiting similar effects.



whereas pharmacological or genetic inhibitors of Lck significantly sup-pressedH/R-induced Sp1 activation, suggesting that Lck and Syk differ-entially regulate the Sp1 activation in response to H/R.

Syk and Lck Differentially Regulate MelCAM, VEGF, and MMP-9Expressions in Response toH/R—Todelineate the roles of Syk and Lck indifferentially controlling the expressions ofMelCAMandVEGF and theactivation ofMMP-9 in the presence of H/R, cells were transfected withDN Lck or WT and Mut Syk and then exposed to H/R. The levels of

MelCAM, VEGF, and MMP-9 were detected by Western blots usingtheir specific antibodies. The results showed that DN Lck and WT Sykbut not Mut Syk restored the H/R-suppressed MelCAM expression(Fig. 4A, lanes 1–5). The levels of Lck and Syk expression were alsodetected by Western blot analysis using their specific antibodies (Fig.4A, 1st and 2nd middle panels). Actin was used as loading control (Fig.4A, lower panel). In contrast, WT Syk and DN Lck suppressed the H/R-induced VEGF expression and MMP-9 activation (Fig. 4, B, lanes 1–4,

FIGURE 2. A and B, role of H/R in regulation of Lck-dependent Syk phosphorylation. A, cells were exposed to hypoxia for 2 h followed by reoxygenation for 0 – 60 min as described. Celllysates were immunoprecipitated with anti-Syk antibody followed by Western blot with anti-phosphotyrosine antibody. The same blot was reprobed with anti-Syk antibody. Notethat significant down-regulation of Syk phosphorylation was observed at 45 min of exposure to H/R. B, cells were pretreated with Lck inhibitors (pp2 (4 nM), aminogenistein (2 �M),and damnacanthal (0.8 �M)) and then exposed to H/R. The level of phosphorylated Syk in the cell lysates was detected by Western blot as described earlier. Note that H/Rdown-regulates Syk phosphorylation, and inhibition of Lck restored the Syk phosphorylation in response to H/R. C, H/R down-regulates Syk-Lck interaction. Cells were either exposedto H/R or individually transfected with WT or Mut Syk and then exposed to H/R. Cell lysates were immunoprecipitated with anti-Lck antibody and analyzed by Western blot (WB) usinganti-Syk antibody. The same blot was reprobed with anti-Lck antibody. All these bands were quantified by densitometric analysis, and the fold changes were calculated. D, role of H/Rin regulation of colocalization between Syk and Lck by immunofluorescence. Note that Syk was stained with TRITC-conjugated IgG, and Lck was stained with FITC-conjugated IgG.The enhanced colocalization was observed at initial stage (control) and restored at 60 min of H/R treatment as indicated by arrows. However, no colocalization was observed at 45 mindue to cytoplasmic translocation of Syk. The results shown here represent three experiments exhibiting similar effects.

FIGURE 3. H/R regulates Syk-Lck-mediated Sp1nuclear translocation and DNA binding. A, cellswere exposed to hypoxia for 2 h followed byreoxygenation for 0 –180 min. The level of Sp1 innuclear and cytoplasmic extracts was detected byWestern blot using anti-Sp1 antibody. Actin wasused as loading control. B, cells were individuallytransfected with WT or Mut Syk or DN Lck and thenexposed to H/R for 60 min as described earlier. TheSp1 level in the nuclear extract was determined byWestern blot. The levels of Syk and Lck in thenuclear extracts were also analyzed by Westernblot using anti-Syk or anti-Lck antibody. The sameblots were reprobed with anti-actin antibody asloading control. C, experiments paralleling that ofA but measuring Sp1-DNA binding by EMSA. D andE, cells were individually transfected with WT andMut Syk or pretreated with pp2 (4 nM) or amin-ogenistein (2 �M) or damnacanthal (0.8 �M) andthen exposed to H/R. The Sp1-DNA binding wasperformed by EMSA. The results shown here repre-sent three experiments exhibiting similar effects.



and C, lanes 1–5). However, cells transfected with Mut Syk furtherenhanced the H/R-induced VEGF expression and MMP-9 activation(Fig. 4, B, lane 3, and C, lane 5). These data suggested that Syk, which isconsidered a negative regulator of tumor progression, can also regulatethe expression of MelCAM, a breast tumor suppressor molecule.

MelCAMNegatively Regulates H/R-induced uPA Secretion andMMP-9Activation—Our previous study demonstrated that uPA plays a crucialrole in MMP-9 activation (47). To examine the role of MelCAM inregulation of uPA secretion and MMP-9 activation, cells were eitherexposed to H/R or pretreated withMelCAM blocking peptide and thenexposed to H/R. Cell lysates were analyzed by Western blot using anti-uPA antibody. Similarly, the conditioned media were used to detect theMMP-9 activation by zymography. The results clearly indicated thatH/R enhances uPA secretion and MMP-9 activation, which is furtherup-regulated in the presence of MelCAM blocking peptide (Fig. 4, D,lanes 1–3, and E, lanes 1–3). Our findings clearly demonstrated that amolecular link exists between Syk/Lck and MelCAM and betweenVEGF and MMP-9, and all of these ultimately control the tumor pro-gression and angiogenesis.

Effect of Syk onXenograftTumorGrowth—Our in vitro results promptedus to extend our studies to the in vivo system. Accordingly, MCF-7 cellswere transfectedwithWTorMutSyk, exposed toH/R, and then implantedinto the nude mice. After 4 weeks, mice were sacrificed, and tumor speci-mens were excised. A portion of tumor samples was used for histopatho-logical studies. The histopathological analysis data indicated thatMut Syk-transfected tumors showed higher infiltration toward the extravasations,nuclear polymorphism, andmitotic features compared with the control orWTSyk-transfected tumors (Fig. 5A,panelsa–c).The immunohistochem-ical studies were also performed with a mixture of anti-Syk and anti-Lckantibodies or with anti-MelCAM, anti-uPA, anti-MMP-9, and anti-vWFantibodies. The data indicated that tumors generated by injectingWTSyk-transfected cells revealed higher nuclear localization and enhanced inter-action between Syk and Lck, whereasMut Syk-transfected tumors showedenhanced cytoplasmic localization of Syk (Fig. 5A, panels d–f). Most inter-

estingly, the expressions of MMP-9, uPA, and neovascularization (vWFexpression) were also higher in the Mut Syk-transfected tumor (Fig. 5E,panels a–i) that further correlated with our in vitro study. As expected, theMelCAM expression was higher in WT Syk-transfected tumor (Fig. 5A,panels g–i).Thiswas also confirmedbyWesternblot analysis (Fig. 5C, lanes1–3).Western blot data also suggested that the levels of uPA andVEGF arehigher in Mut Syk-transfected tumors (Fig. 5B, lanes 1–3). Similar resultswere obtained in Sp1-DNA binding as detected by EMSA (Fig. 5D, lanes1–3). Thus our studies clearly indicated that Syk acts as a tumor suppressorand anti-angiogenic candidate gene that differentially regulates MelCAMexpression thatultimately controls uPA-dependentMMP-9activation. Sykalso negatively regulates VEGF expression. These signalingmolecules ulti-mately control the tumor progression and angiogenesis.

Human Breast Tumor Specimen Analysis—The in vitro and in vivomouse model data further prompted us to extend these studies withhuman clinical breast tumor samples. Human solid breast tumor spec-imens were collected with the informed consent from a local hospital.The tumor grading was determined by histopathology analysis byhematoxylin and eosin staining using amodified Scarff-Bloom-Richard-son system, and photographs were taken with a Nikonmicroscope (Fig.6A, panels a–d). The cellular localizations of Syk and Lck in humanclinical breast tumor sections were detected by incubating with a mix-ture of anti-Syk and anti-Lck antibodies. Our results indicated thatin normal breast tissue there was colocalization of Syk and Lck, butwith the increasing grade of tumor specimens, no colocalization wasobserved. Furthermore, in higher grades of tumors, the expression ofSyk was totally diminished (Fig. 6A, panels e–h). As expected,MelCAMexpression was significantly lower in higher grades of tumors comparedwith lower grades or normal tissue (Fig. 6A, panels i–l). The expressionof HIF-1� was detected byWestern blot using the same specimens, andthe data showed that the level of HIF-1� was significantly higher ingrade II and III tumors (Fig. 6B, lanes 1–4), which indicated the hypoxicstatus in the tumor microenvironment. Higher expressions of MMP-9,uPA, VEGF, and vWF in these tumor specimens further correlated with

FIGURE 4. Syk-Lck interplay differentially regu-lates H/R-induced expression of downstreameffector molecules. A, cells were individuallytransfected with WT and Mut Syk or DN Lck andexposed to H/R. The level of MelCAM expression inthe cell lysates was analyzed by Western blot. Thelevels of Syk and Lck were analyzed by Westernblot. The same blots were reprobed with anti-actinantibody as loading control. B and C, the expres-sion of VEGF in the above transfected cell lysatesand the activation of MMP-9 in the above condi-tioned media were detected by Western blotusing their specific antibodies. Human recombi-nant VEGF was used as positive control. D and E,cells were pretreated with MelCAM blocking pep-tide and then exposed to H/R. The cell lysates andconditioned media were used to detect the levelsof uPA by Western blot (D) and MMP-9 by zymog-raphy (E), respectively. Note that H/R down-regu-lates MelCAM expression but enhances VEGFexpression and MMP-9 activation. Blocking of Mel-CAM by its blocking peptide further stimulates H/R-induced uPA expression and MMP-9 activation.



FIGURE 5. H/R regulates tumor growth and angiogenesis through differential interplay between Syk and Lck. A, MCF-7 cells (5 � 106/0.2 ml) were either exposed to H/R for 24 hor transfected with WT or Mut Syk and then exposed to H/R. The cell suspension was mixed with cold Matrigel and injected subcutaneously in the dorsal region of mice. After 4 weeks,tumors were excised and subjected to histopathological analysis by hematoxylin and eosin staining (panels a– c). Note that increased infiltration and nuclear polymorphism wereobserved in a tumor generated by injecting Mut Syk but not WT Syk-transfected cells. The physical association between Syk and Lck in these tumors was detected by immunohis-tochemistry (panels d–f). Note that tumors generated by injecting WT Syk-transfected cells showed higher nuclear localization and enhanced interaction between Syk and Lck,whereas Mut Syk-transfected tumors showed enhanced cytoplasmic localization of Syk (panels d–f). Syk was stained with FITC-conjugated IgG (green), and Lck was stained withTRITC-conjugated IgG (red). The nuclei of the cells was visualized by DAPI (blue) staining. The MelCAM expression in these tumors was analyzed by immunofluorescence usinganti-MelCAM antibody (panels g–i). B and C, the expressions of uPA, VEGF, and MelCAM in the tumor extracts were also analyzed by Western blot using their specific antibodies. D, theSp1-DNA binding in the nuclear extracts was performed by EMSA. E, the levels of uPA, MMP-9, and vWF in the above tumor specimens were analyzed by immunofluorescence usingtheir specific antibodies (panels a–i). Note that there was drastic reduction in uPA and MMP-9 levels in a tumor generated by injecting WT Syk but not Mut Syk-transfected cells. Themicrovessel density was detected by staining with anti-vWF antibody (endothelial cell-specific marker) in these tumors (panels g–i). The nuclei were stained with DAPI (blue). Theenhanced vWF expression observed in Mut Syk-transfected tumors indicated higher vascularization compared with control or WT Syk-transfected tumors.



higher grades (Fig. 6C, panels a–p). Taken together, our data indicatedthat the Syk andMelCAM act as a negative regulator of tumor progres-sion and angiogenesis and can act as a potential therapeutic target inbreast cancer.

DISCUSSION

Our study indicated that Syk and Lck play a crucial regulatory factorin hypoxia-induced tumor progression and angiogenesis. The data alsodemonstrated that Syk acts as a negative regulator of H/R-induced

tumor progression, whereas Lck as a positive regulator, and H/R plays apivotal role in regulating the cross-talk between Syk and Lck. Our pre-vious study indicated that inhibition of Syk activity by Syk-specific anti-sense S-oligonucleotide resulted in enhanced uPA expression and cellmotility in MCF-7 cells (23). In this study, we showed that H/R plays acrucial role in inactivation of Syk by inhibiting its phosphorylation,thereby resulting in its cytoplasmic translocation. Furthermore, the dataindicated that the physical association between phosphorylated Syk andLck in the nucleus resulted in inactivation of Lck. Upon H/R treatment

FIGURE 6. Immunohistochemical analysis ofhuman breast tumor specimens of differentpathological grades. A, normal and tumorigenichuman breast specimens were collected from alocal hospital with informed consent, and the gra-dations of these samples were performed using amodified Scarff-Bloom-Richardson system (panelsa– d). Expression and cellular colocalization of Sykand Lck were analyzed by immunohistochemistry(panels e– h). Syk was stained with FITC, and Lckwas stained with TRITC, and colocalization wasvisualized as yellow. The nuclei were stained withDAPI (blue). Note that in normal breast tissue thecellular colocalization of Syk-Lck was observed.The expression of both Syk and Lck were observedin grade I tumor. However, Lck but not Syk expres-sion was observed in tumors of higher grades(grades II and III). There was significant expressionof MelCAM in normal and lower grade tumors butwas significantly diminished in tumors of highergrades (panels i–l). B, hypoxic level of these speci-mens was determined from the tumor lysates byWestern blot analysis using anti-HIF-1� antibody.Actin was used as loading control. C, expressionlevels of MMP-9 (panels a– d), uPA (panels e– h),VEGF (panels i–l), and vWF (panels m–p) were ana-lyzed by immunohistochemistry using their spe-cific antibodies. Note that the expression status ofall of these molecules was significantly enhancedin higher grades of tumors.



the phosphorylation of Syk was abrogated, which in turn resulted in thedissociation of Syk from Lck leading to activation of Lck, which ulti-mately targets downstream signaling events.Previous studies indicated thatMelCAMact as a tumor suppresser in

breast carcinoma (32). This prompted us to investigate whether H/Rregulates Syk/Lck-mediated MelCAM expression in MCF-7 cells. Ourdata indicated that H/R down-regulates the expression of MelCAM,which was restored upon transfection of cells with WT Syk or DN Lck.Our findings established a crucialmolecular link between the two breastcancer-specific tumor suppressor molecules, Syk and MelCAM. Thedata clearly indicated that Syk enhances and Lck inhibits MelCAMexpression, and thus both ultimately regulate breast cancer progression.Moreover, H/R-induced expressions of MMP-9 and VEGF were sup-pressed upon overexpression of Syk or inactivation of Lck. Thus Sykexerts its tumor-suppressive effect either through induction of Mel-CAM expression or suppression of tumor promoters like uPA,MMP-9,and VEGF. On the other hand, Lck activated upon H/R treatment,which enhances breast tumor progression through suppression of Mel-CAM expression and induction ofMMP-9, uPA, and VEGF expression.Interestingly, the data also showed that blocking of MelCAM using itsspecific blocking peptide also resulted in induction of uPA as well asMMP-9 expression and activation. The data further indicated that Syksuppresses tumor progression through regulation of MelCAM-medi-ated mechanism.Sp1 is a ubiquitously expressed transcription factor that recognizes

GC-rich sequences present in the regulatory sequences of numeroushousekeeping genes and those genes that are involved in growth regu-lation and cancer (34, 45). The transcription factor Sp1-response ele-ment is present in the promoter region of various genes, including uPA,MMP-9, and VEGF. Sp1 is one of the key transcription factors, and itsactivation occurs under hypoxic conditions in various cancer cells,including breast carcinoma (34, 36, 45). In our study, we have demon-strated that H/R-regulated cross-talk between Syk and Lck controlsnuclear localization and DNA binding of Sp1. Furthermore, Syk nega-tively and Lck positively regulates Sp1 activation. Therefore, H/Rthrough a Syk-Lck-mediated pathway controls Sp1 activation and reg-ulates the expression of downstream molecules such as uPA, MMP-9,

VEGF, and MelCAM that promotes tumor growth and angiogenesis(Fig. 7).Our in vivo and clinical specimen analysis clearly supports our

in vitro findings. In a xenograft study, tumors generated by injectingwild type Syk-transfected cells significantly suppressed the tumorgrowth, whereas mutant Syk-transfected cells showed enhanced tumorgrowth. The levels of uPA, MMP-9, VEGF, as well as microvessel den-sity were significantly reduced, whereas the physical interactionbetween Syk and Lck and expression of MelCAM were significantlyenhanced inwild type but not inMut Syk-transfected tumors. In clinicalspecimen analysis, we observed that in higher grades of tumors (grade IIand III), the level of HIF-1� is significantly higher, which indicated theenhanced hypoxic condition in tumormicroenvironment. The low levelof HIF-1� was observed in normal breast tissue and in lower grades oftumor. There was significant expression of Syk in normal breast tissuesand lower grades of tumor. However, the physical association betweenSyk and Lck was observed in normal breast tissues. Moreover, theincreased expression of Lck but not Syk was visualized in higher gradesof tumors. Furthermore, the higher grades of tumors showed enhancedexpression of uPA, MMP-9, and VEGF, and the increased microvesseldensity was also characterized in these tumors. In addition, the Mel-CAM expression was significantly diminished in tumors of highergrades that corroborated with both in vitro and in vivo findings.In summary, to our knowledge, this is the first report that H/R differ-

entially regulates the cross-talk between Lck and Syk, which ultimatelycontrol tumor progression. Previous reports have indicated that anincreased hypoxic or consecutive hypoxia/reoxygenation condition inthe tumor microenvironment plays a crucial role in determining theoncogenic potential of various cancers. Therefore, an in-depth under-standing of theH/R-regulated signalingmechanismmay be beneficial indesigning a novel therapeutic approach for the treatment of cancer.Wehave shown that Syk, a tumor suppressor, positively regulates theexpression of the breast cancer-specific tumor suppressor MelCAM.Interestingly, overexpression of Syk inhibits H/R-induced expression ofseveral tumorigenicmolecules, including uPA,MMP-9, andVEGF.Ourresults indicate that the mechanism demonstrated in the mouse modelunderlies human pathology, and a clear understanding of such mecha-

FIGURE 7. Molecular mechanism of H/R-regu-lated cross-talk between Syk and Lck that con-trols breast tumor progression and angiogene-sis. H/R suppresses Syk phosphorylation leadingto its cytoplasmic translocation. This results in acti-vation of Lck, which initiates the various down-stream signaling events. Lck in response to H/R stim-ulates Sp1 activation leading to enhanced uPA,MMP-9, and VEGF expressions but suppression ofMelCAM expression, which ultimately stimulatesbreast carcinoma progression and angiogenesis.



nisms may facilitate the development of novel therapeutic approachesto suppress hypoxia/reoxygenation-regulated Syk/Lck-mediated uPA,MMP-9, and VEGF expression, thereby controlling tumor growth andangiogenesis.

REFERENCES1. Hockel, M., and Vaupel, P. (2001) J. Natl. Cancer Inst. 93, 266–2762. Bussink, J., Kaanders, J. H., and van der Kogel, A. J. (2003) Radiother. Oncol. 67, 3–153. Dachs, G. U., and Tozer, G. M. (2000) Eur. J. Cancer 36, 1649–16604. Brennan, P. A., Mackenzie, N., and Quintero, M. (2005) J. Oral. Pathol. & Med. 34,

385–3895. Liu, L., and Simon, M. C. (2004) Cancer Biol. Ther. 3, 492–4976. Vaupel, P., Kallinowski, F., and Okunieff, P. (1989) Cancer Res. 49, 6449–64657. Vaupel, P., Thews, O., Kelleher, D. K., and Hoeckel, M. (1998) Adv. Exp. Med. Biol.

454, 591–6028. Hockel, M., Schlenger, K., Hockel, S., and Vaupel, P. (1999) Cancer Res. 59,

4525–45289. Koong, A. C., Denko, N. C., Hudson, K.M., Schindler, C., Swiersz, L., Koch, C., Evans,

S., Ibrahim, H., Le, Q. T., Terris, D. J., and Giaccia, A. J. (2000) Cancer Res. 60,883–887

10. Fan, C., Li, Q., Ross, D., and Engelhardt, J. F. (2003) J. Biol. Chem. 278, 2072–208011. Dachs, G. U., Patterson, A. V., Firth, J. D., Ratcliffe, P. J., Townsend, K. M., Stratford,

I. J., and Harris, A. L. (1997) Nat. Med. 3, 515–52012. Chaplin, D. J., Hill, S. A., Bell, K. M., and Tozer, G. M. (1998) Semin. Radiat. Oncol. 8,

151–16313. Kunz, M., and Ibrahim, S. M. (2003)Mol. Cancer 2, 1–1314. Kunz, M., Bloss, G., Gillitzer, R., Gross, G., Goebeler, M., Rapp, U. R., and Ludwig, S.

(2002) Biochem. J. 366, 299–30615. Natarajan, R., Jones, D. G., Fisher, B. J., Wallace, T. J., Ghosh, S., and Fowler, A. A.

(2005) Biochem. Cell Biol. 83, 597–60716. Said, H. M., Katzer, A., Flentje, M., and Vordermark, D. (2005) Radiother. Oncol. 76,

200–20517. Kurosaki, T. (2000) Curr. Opin. Immunol. 12, 276–28118. Turner,M., Schweighoffer, E., Colucci, F., Di Santo, J. P., and Tybulewicz, V. L. (2000)

Immunol. Today 21, 148–15419. Yanagi, S., Kurosaki, T., and Yamamura, H. (1995) Cell. Signal. 7, 185–19320. Cheng, A. M., Negishi, I., Anderson, S. J., Chan, A. C., Bolen, J., Loh, D. Y., and

Pawson, T. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 9797–980121. Ulanova, M., Puttagunta, L., Marcet-Palacios, M., Duszyk, M., Steinhoff, U., Duta, F.,

Kim, M. K., Indik, Z. K., Schreiber, A. D., and Befus, A. D. (2005) Am. J. Physiol. 288,497–507

22. Coopman, P. J., Do,M. T., Barth, M., Bowden, E. T., Hayes, A. J., Basyuk, E., Blancato,J. K., Vezza, P. R., McLeskey, S. W., Mangeat, P. H., and Mueller, S. C. (2000) Nature406, 742–747

23. Mahabeleshwar, G. H., and Kundu, G. C. (2003) J. Biol. Chem. 278, 6209–622124. Koster, A., Landgraf, S., Leipold, A., Sachse, R., Gebhart, E., Tulushan, A. H., Ronay,

G., Schmidt, C., and Dingermann, T. (1991) Anticancer Res. 11, 193–20125. Veillette, A., Bookman, M. A., Horak, E. M., and Bolen, J. B. (1988) Cell 55, 301–30826. Couture, C., Deckert, M., Williams, S., Russo, F. O., Altman, A., and Mustelin, T.

(1996) J. Biol. Chem. 271, 24294–2429927. Lehmann, J.M., Holzmann, B., Breitbart, E.W., Schmiegelow, P., Riethmuller, G., and

Johnson, J. P. (1987) Cancer Res. 47, 841–84528. Lehmann, J. M., Riethmuller, G., and Johnson, J. P. (1989) Proc. Natl. Acad. Sci.

U. S. A. 86, 9891–989529. Sers, C., Kirsch, K., Rothbacher, U., Riethmuller, G., and Johnson, J. P. (1993) Proc.

Natl. Acad. Sci. U. S. A. 90, 8514–851830. Xie, S., Luca, M., Huang, S., Gutman, M., Reich, R., Johnson, J. P., and Bar-Eli, M.

(1997) Cancer Res. 57, 2295–230331. Rummel, M. M., Sers, C., and Johnson J. P. (1996) Cancer Res. 56, 2218–222332. Shih, L. M., Hsu, M. Y., Palazzo, J. P., and Herlyn, M. (1997) Am. J. Pathol. 151,

745–75133. Shih, I. M. (1999) J. Pathol. 189, 4–1134. Karlen, S., and Braathen, L. R. (1999) J. Investig. Dermatol. 113, 711–71935. Collen, D. (1999) Thromb. Haemostasis 82, 259–27036. Blasi, F. (1999) Thromb. Haemostasis 82, 298–30437. Shapiro, S. D. (1998) Curr. Opin. Cell Biol. 10, 602–60838. Folkman, J. (1999) Nat. Biotechnol. 17, 74939. Rangaswami, H., Bulbule, A., and Kundu, G. C. (2004) J. Biol. Chem. 279,

38921–3893540. Ferrara, N., Gerber, H. P., and LeCouter, J. (2003) Nat. Med. 9, 669–67641. Neufeld, G., Cohen, T., Gengrinovitch, S., and Poltorak, Z. (1999) FASEB J. 13, 9–2242. Lutsenko, S. V., Kiselev, S. M., and Severin, S. E. (2003) Biochemistry (Mosc.) 68,

286–30043. Matsumoto, T., and Claesson-Welsh, L. (2001) Sci. STKE. 2001 RE2144. Mercurio, A. M., Bachelder, R. E., Bates, R. C., and Chung, J. (2004) Semin. Cancer

Biol. 14, 115–12245. Bradbury, D., Clarke, D., Seedhouse, C., Corbett, L., Stocks, J., and Knox, A. (2005)

J. Biol. Chem. 280, 29993–3000046. Philip, S., Bulbule, A., and Kundu, G. C. (2001) J. Biol. Chem. 276, 44926–4493547. Rangaswami, H., Bulbule, A., and Kundu, G. C. (2005) J. Biol. Chem. 280,

19381–1939248. Sounni, N. E., Devy, L., Hajitou, A., Frankenne, F., Munaut, C., Gilles, C., Deroanne,

C., Thompson, E. W., Foidart, J. M., and Noel, A. (2002) FASEB J. 16, 555–564



Osteopontin Promotes Vascular Endothelial Growth

Factor–Dependent Breast Tumor Growth and Angiogenesis

via Autocrine and Paracrine Mechanisms

Goutam Chakraborty, Shalini Jain, and Gopal C. Kundu


Abstract

Angiogenesis is the hallmark of cancer, and developmentof aggressiveness of primary tumor depends on de novoangiogenesis. Here, using multiple in vitro and in vivo models,we report that osteopontin (OPN) triggers vascular endothe-lial growth factor (VEGF)–dependent tumor progression andangiogenesis by activating breast tumor kinase (Brk)/nuclearfactor–inducing kinase/nuclear factor-KB (NF-KB)/activatingtranscription factor-4 (ATF-4) signaling cascades throughautocrine and paracrine mechanisms in breast cancer system.Our results revealed that both exogenous and tumor-derivedOPN play significant roles in VEGF-dependent tumor angio-genesis. Clinical specimen analysis showed that OPN andVEGF expressions correlate with levels of neuropilin-1, Brk,NF-KB, and ATF-4 in different grades of breast cancer.Consequently, OPN plays essential role in two key aspects oftumor progression: VEGF expression by tumor cells and VEGF-stimulated neovascularization. Thus, targeting OPN and itsregulated signaling network could be a novel strategy to blocktumor angiogenesis and may develop an effective therapeuticapproach for the management of breast cancer. [Cancer Res2008;68(1):152–61]

Introduction

Angiogenesis, the formation of new blood vessels from theexisting one is a key step for tumor growth, survival, progression,and metastasis (1). A large number of proangiogenic factors andtheir cognate receptors have been identified (2–4). To date, the bestcharacterized proangiogenic cytokine to turn on the ‘‘angiogenicswitch’’ is vascular endothelial growth factor (VEGF; ref. 5). Forover a decade, the role of VEGF in regulation of tumor angiogenesishas been under intense investigation (6). Previous results indicatedthat VEGF signaling represents a critical rate-limiting step inangiogenesis (7). Alternative exon splicing results in four isoformsof VEGF (VEGF121, VEGF165, VEGF189, and VEGF206), and amongthem, VEGF165 is the predominant one that plays a major role intumor angiogenesis (7, 8). Recent data show that the functions ofVEGF may not be limited to endothelial cells but also playimportant roles in survival, proliferation, and migration in tumorcells (9, 10).Osteopontin (OPN), a secreted noncollagenous, sialic acid–rich,

chemokine-like protein and also a member of small integrin-

binding ligand N-linked glycoprotein family plays important role indetermining the oncogenic potential of various cancers (11, 12).OPN exerts its prometastatic effects by regulating various cellsignaling events through interaction with integrins and CD44receptors that ultimately lead to tumor progression (13). Recentevidences indicated that OPN regulates tumor growth throughinduction of cyclooxygenase-2 and urokinase-type plasminogenactivator expressions and activation of matrix metalloproteinasesin various cancer cells (14–17). The role of OPN in variouspathophysiologic conditions, particularly in cancer, suggested thatthe variation in glycosylation, phosphorylation, and sulfationgenerates the different functional forms that might alter its normalphysiologic functions (12, 18). Previous studies have shown thattumor-derived OPN is soluble, and it exhibits close similarity withhuman milk OPN (19, 20). Earlier reports also suggested that OPNproduced either from tumor or stromal cells has been shown toenhance the metastatic ability (21).Breast tumor kinase (Brk/PTK6/Sik) is a nonreceptor tyrosine

kinase and expressed in metastatic breast tumors but not in normalmammary tissue or benign lesions (22). It has been postulated thatBrk promotes cell migration, and silencing of Brk resulted ininhibition of migration and proliferation in various cancer cells,includingMDA-MB-231 (23, 24). Recent data revealed that activatingtranscription factor-4 (ATF-4/cAMP-responsive element bindingprotein-2) regulates VEGF expression (25, 26). Neuropilin-1 (NRP-1)was initially characterized to mediate the chemorepulsive activityof collapsin/semaphorins in neuronal cells (27, 28). NRP-1 is alsoexpressed in endothelial cells and in various tumors, and it functionsas isoform-specific receptor for VEGF165 and plays an important rolein tumor progression (29–32). Previous studies have shown thatblocking or silencing of NRP-1 resulted in significant reduction oftumor progression (10). However, the role of Brk in regulation ofATF-4–dependent VEGF expression and interaction of VEGF withNRP-1 in response to OPN and how all of these ultimately controlbreast tumor angiogenesis is not well defined.In this study, we provide both in vitro and in vivo experimental

evidences that show the molecular mechanism by which OPNregulates Brk/nuclear factor-nB (NF-nB)/ATF-4 signaling cascadesthat ultimately augment the VEGF expression and tumorangiogenesis through autocrine and paracrine mechanisms. Wehave further substantiated the crucial roles of tumor-derived,endogenous OPN in tumor angiogenesis using short interferingRNA (siRNA) based approach in in vitro and in vivo models.Moreover, our clinical data revealed that the enhanced expressionsof OPN and VEGF correlate with NRP-1, Brk, NF-nB and ATF-4levels in breast carcinoma of higher grades. These data provide newinsights into the mechanism underlying the regulation of VEGFexpression by OPN in breast tumor angiogenesis and understand-ing these mechanisms may form the basis of new therapeuticregimens for the management of breast cancer.

Note: Supplementary data for this article are available at Cancer Research Online(http://cancerres.aacrjournals.org/).

Requests for reprints: Gopal C. Kundu, National Center for Cell Science, Pune 411007, India. Phone: 91-20-25708103; Fax: 91-20-25692259; E-mail: [email protected].


Cancer Res 2008; 68: (1). January 1, 2008 152 www.aacrjournals.org

Research Article


Cell culture and transfection. Human breast adenocarcinoma (MDA-MB-231) cell line was obtained from American Type Culture Collection.Human umbilical vein endothelial cell line EA.hy-926 was a generous giftfrom Dr. Christopher Newton. The super repressor form of InBa in pCMV4(Dr. Dean Ballard, Vanderbilt University School of Medicine), wild-type (wt)and kinase negative (mut) nuclear factor–inducing kinase (NIK; K249A/K430A) in pcDNA3 vector (Prof. David Wallach, Weizmann Institute ofScience), VEGF-Luc construct (Dr. Debabrata Mukhopadhyay, Mayo ClinicCollege of Medicine), VEGF-GFP cDNA construct (Dr. Rakesh K. Jain,Massachusetts General Hospital), wt (pEF/mATF-4) and dominant-negative(dn; pEF/mATF-4M) ATF-4 (Dr. Javed Alam, Yale University School ofMedicine), wt and kinase mutant (KM) Brk in pRC CMV vector (Prof. MarkCrompton, Royal Holloway University of London), and human OPN cDNA inpcDNA 3.1 (Dr. Ann Chambers, University of Western Ontario) weretransfected in MDA-MB-231 cells using Lipofectamine 2000.siRNA. Cells were transfected with OPN siRNA (OPNi) or NRP-1 siRNA

(NRP-1i; Dharmacon) using liposome according to manufacturer’s instruc-tions. The sequence targeted for OPN is 5¶ GUU UCA CAG CCA CAA GGACdTdT/dTdT CAA AGU GUC GGU GUU CCU G 5¶ and its nonsilencingcontrol is 5¶ CAG UAC AAC GCA UCU GGC AdTdT/dTdT GUC AUG UUGCGU AGA CCG U 5¶. The NRP-1 siRNA sequence is 5¶ GAG AGG UCC UGAAUG UUC CdTdT/dTdT CUC UCC AGG ACU UAC AAG G 5¶ and its non-silencing control is 5¶ AGA GAU GUA GUC GCU CGC UdTdT/dTdT UCUCUA CAU CAG CGA GCG A 5¶. The Brk siRNA sequence is 5¶AAG GUG GCCAUU AAG GUG AdTdT/dTdT UUC CAC CGG UAA UUC CAC U 5¶ and itsnonsilencing control is 5¶CAC ACU AGG UUG CCA CAG GdTdT/dTdT GUGUGA UCC AAC GGU GUC C 5¶.Purification of human OPN. The human OPN was purified from human

breast milk as described previously with minor modifications and usedthroughout this study (15).Coculture assay. The tumor-endothelial cell interaction was studied by

coculture experiments using MDA-MB-231 and EA.hy-926 cells. EA.hy-926cells were transfected with NRP-1i and cocultured with nontransfected orOPNi or NRP-1i transfected MDA-MB-231 cells. In separate experiments,cocultured cells were either treated with OPN alone or along with anti-VEGF antibody. The level of pKDR in cell lysates was detected byimmunoprecipitation followed by Western blot.

Immunoprecipitation and Western blot. The immunoprecipitation

and immunoblot experiments were performed using their specific anti-bodies as described (14).

RNA extraction and reverse transcription–PCR. The RNA wasisolated from MDA-MB-231 cells and used for reverse transcription–PCR(RT-PCR). The VEGF primers were used: sense, 5¶-CCC TCC GAA ACC ATGAAC TTT-3¶ and antisense, 5¶-AGA GAT CTG GTT CCC GAA AC-3¶. Theamplified cDNA fragments were resolved by agarose gel electrophoresis.

Reporter gene expression. MDA-MB-231 cells were transfected withVEGF-luciferase reporter construct, either alone or along with OPN cDNAor OPNi. Cell lysates were analyzed by Western blot using antiluciferaseantibody. The VEGF promoter activity was measured either by transfectingMDA-MB-231 cells with VEGF-GFP or cotransfected with wt and KM Brk orwt and dn ATF-4 or InBa super repressor (sup. rep.) or Brk siRNA and thentreated with OPN. The GFP expression was measured by using fluorescencemicroscopy (Nikon) and quantified by Image Pro Plus 6.0 Software (Nikon)and represented in the form of bar graph.Brk kinase assay. Brk kinase assay was performed as described earlier (22).EMSA. MDA-MB-231 cells were either transfected with wt and KM Brk

or wt and mut NIK, InBa sup. rep. or wt and dn ATF-4 and then treatedwith 0.5 Amol/L OPN. Nuclear extracts were used for NF-nB and ATF-4–DNA binding.

Immunofluorescence. To examine whether OPN regulates ATF-4expression and nuclear translocation, MDA-MB-231 cells were treated with

0.5 Amol/L OPN, and immunofluorescence was performed using anti–ATF-4

antibody.

Wound migration. The wound assay was performed using MDA-MB-231and endothelial (EA.hy-926) cells with typical cobblestone morphology asdescribed (14).

Cell comigration assay. The endothelial-breast tumor cell interactionwas shown by comigration assays. MDA-MB-231 cells were treated with

either OPN alone or along with anti-VEGF or anti–NRP-1 blocking antibody

and used in lower chamber. EA.hy-926 cells were seeded on the upper

chamber. In separate experiments, MDA-MD-231 cells were transfectedwith OPNi or NRP-1i and used for comigration assay. The migrated cells

were stained with Giemsa and were counted in three high-power fields

under an inverted microscope.

In vivo Matrigel-based angiogenesis assay. The in vivo Matrigelangiogenesis assay was preformed as described (33). Briefly, Matrigel

(0.5 mL) was injected s.c. in the ventral groin region of female athymic

NMRI (nu/nu) mice. In separate experiments, the conditioned medium

(CM) of untreated or OPN-treated MDA-MB-231 cells were mixed withMatrigel and injected to the mice. In other experiments, OPN-treated CM

was mixed with anti-VEGF blocking antibody (400 Ag/kg of body weight permice) and Matrigel or CM collected from OPNi-transfected cells were mixedwith Matrigel and then injected to the mice. After 3 weeks, Matrigel plugs

were excised and processed for histopathology and immunohistochemistry.

The paraffin sections were immunostained with antihuman vWF and

anti–phosphorylated KDR antibodies.In vivo tumorigenicity, histopathology, and immunohistochemistry.

The tumorigenicity and immunohistochemistry experiments were per-

formed as described earlier (34, 35). Briefly, MDA-MB-231 (5 � 105) cells

were injected orthotopically in the mammary fat pad of female athymicnude mice. OPN (0.5 Amol/L) or anti-VEGF (400 Ag/kg of body weight permice) blocking antibody was injected to the site of tumor twice a week up to

6 weeks. In other experiments, OPN cDNA transfected cells alone or alongwith anti-VEGF antibody were injected into the nude mice. In separate

experiments, OPNi or NRP-1i (250 Ag/kg of body weight per mice) mixed

with transfection reagent were given to the site of tumor twice a week until

termination of the experiments. Mice were sacrificed and photographed, thetumors were dissected out and weighed, and tumor tissues were used for

histopathologic and immumohistochemical studies with their specific

antibodies. Tumor samples were lysed in lysis buffer. Mice blood was

collected from retro-orbital plexus, and serum was isolated. The levels ofOPN, VEGF, and CA 15-3 were detected by Western blot using their specific

antibodies. The OPN-induced VEGF–NRP-1 interaction in tumor samples

was analyzed by immunoprecipitating tumor lysates with anti-VEGFantibody and immunoblotted with anti–NRP-1 antibody. The nuclear

extracts of tumor tissues were prepared as described earlier and used for

EMSA for detection of NF-nB–DNA and ATF-4–DNA binding (35).

Human breast clinical specimen analysis. Human breast tumorspecimens of different grades and normal breast tissues were collected from

local hospital with informed consent and flash frozen. The levels of OPN

and VEGF were examined by Western blot. The tumor sections were stained

with H&E. The vWF expression and microvessel density were detected byimmunofluorescence using anti-vWF antibody. To examine the status of NF-

nB localization and NRP-1, Brk, and ATF-4 expressions, tumor tissue

sections were analyzed by immunohistochemistry using their specific

antibodies and were visualized under confocal microscopy (Zeiss).Statistical analysis. The results of the experimental studies are

expressed as mean F SE. Statistical differences were analyzed by Student’s

t test. P < 0.05 was regarded as significant. All these bands were analyzeddensitometrically (Kodak Digital Science), and the fold changes were

calculated. The in vivo angiogenesis and all other in vivo experimental data

were quantified using the Image Pro Plus 6.0 Software (Nikon).

Results

OPN augments the expression of VEGF. To examine whetherOPN regulates VEGF expression and control breast tumorangiogenesis, MDA-MB-231 cells were treated with exogenousOPN, and VEGF expression were analyzed by Western blot (Fig. 1A)and by RT-PCR (Fig. 1B). The results indicated that OPN inducesVEGF expression both at protein and RNA levels. To examine thespecificity of OPN on VEGF expression, cells were individually

OPN Regulates VEGF-Dependent Tumor Angiogenesis

www.aacrjournals.org 153 Cancer Res 2008; 68: (1). January 1, 2008

transfected with OPN cDNA or OPNi or treated with anti-OPNblocking antibody (R&D Systems). The data revealed that cellstransfected with OPN cDNA enhanced, whereas transfected withOPNi or treated with anti-OPN antibody suppressed, the VEGFexpression (Fig. 1C). To examine whether OPN regulates VEGFpromoter activity, cells were individually transfected with OPNcDNA or OPNi along with VEGF-luciferase reporter construct.Cell lysates were analyzed by Western blot using antiluciferaseantibody. The data showed that silencing of OPN drasticallysuppressed, whereas overexpression significantly enhanced, VEGFpromoter activity (Fig. 1D). These results suggested that exogenous,as well as tumor-derived, OPN augments VEGF expression both attranscriptional and protein levels.OPN stimulates Brk/NIK-dependent NF-KB activation. Brk is

considered as one of the key regulator during breast cancerprogression and is expressed moderately in MDA-MB-231 cells(22–24). We sought to determine whether OPN regulates Brkphosphorylation in MDA-MB-231 cells. Accordingly, cells weretreated with OPN for 0 to 24 min, and level of pBrk was detected byimmunoprecipitation followed by Western blot. The results showedthat OPN enhances Brk phosphorylation (Supplementary Fig. S1A).Pretreatment of cells with avh3 blocking antibody significantlysuppressed, whereas inhibitor of c-Src (pp2) or phosphatidylino-sitol 3-kinase (PI3K; wortmannin) had no effect on OPN-inducedBrk phosphorylation and Brk kinase activation (Fig. 2A). The dataare analyzed densitometrically and represented in the form of bargraph (Fig. 2A). Cells treated with RGD but not with RGE peptidealso suppressed OPN-induced Brk phosphorylation (data notshown). These data suggested that OPN regulates avh3-integrin–dependent Brk activation, which is independent of c-Src/PI3K

signaling pathway. Earlier, we have shown that NIK plays a crucialrole in OPN-induced signaling, which in turn regulates tumorgrowth (18). To examine whether Brk plays any role on OPN-induced NIK phosphorylation, cells were individually transfectedwith wt or KM Brk and then treated with OPN. The data revealedthat wt Brk enhanced, whereas KM Brk inhibited, OPN-inducedNIK phosphorylation, indicating that Brk plays a crucial role in thisprocess (Fig. 2B). EMSA results revealed that wt Brk enhanced,whereas KM Brk suppressed, OPN-induced NF-nB–DNA binding(Supplementary Fig. S1B). Moreover, Brk siRNA inhibited OPN-induced NF-nB–DNA binding (data not shown). These resultsdefine a novel-signaling pathway by which OPN regulates Brk-dependent NIK-mediated NF-nB activation.OPN regulates crosstalk between NF-KB and ATF-4 leading

to VEGF expression. Previous reports have indicated that ATF-4regulates VEGF expression in response to various stimuli (26). Toexamine whether OPN regulates ATF-4 cellular localization, cellswere treated with OPN or transfected with OPNi, and localizationof ATF-4 was detected by immunofluorescence (SupplementaryFig. S1C). The results showed that OPN enhanced ATF-4 nuclearlocalization. To check whether OPN regulates ATF-4–DNA bindingand involvement of Brk in this process, EMSA was performed. Thedata revealed that wt Brk enhanced, whereas KM Brk drasticallysuppressed, OPN-induced ATF-4–DNA binding (Fig. 2C, lanes 1–4).Brk siRNA also suppressed OPN-induced ATF-4–DNA binding(data not shown). InBa sup. rep. or mut NIK inhibited, whereas wtNIK significantly enhanced, OPN-induced ATF-4–DNA binding(Fig. 2C, lanes 5–9). Cotransfection of cells with wt Brk along withInBa sup. rep. or mut NIK unaffected OPN-induced ATF-4–DNAbinding (Fig. 2C, lanes 10 and 12). Similar results were obtained

Figure 1. Exogenous and tumor-derived OPN augments VEGF expression in MDA-MD-231 cells. A, serum-starved cells were incubated with 0 to 1 Amol/L OPNfor 6 h at 37jC, and conditioned media and cell lysates were analyzed by Western blot using anti-VEGF antibody. Actin was used as loading control. B, cells weretreated with 0 to 1 Amol/L OPN for 3 h, total RNA was isolated, and the level of VEGF mRNA was detected by semiquantitative RT-PCR. Glyceraldehyde-3-phosphatedehydrogenase (GAPDH ) was used as internal control. C, cells were transfected with OPN cDNA, OPNi, or Coni or treated with OPN blocking antibody. Theexpressions of VEGF and OPN in cell lysates were analyzed by Western blot using their specific antibodies. Human recombinant VEGF and purified OPN were used ascontrols. D, cells were transfected with VEGF-luciferase reporter construct alone or cotransfected with OPNi, Coni, or OPN cDNA, and luciferase expression wasdetected by Western blot using antiluciferase antibody. The data represent three experiments exhibiting similar results. Fold changes were calculated.

Cancer Research


when cells were cotransfected with KM Brk along with wt NIK(Fig. 2C, lane 11). These data indicated that OPN up-regulates Brk-dependent ATF-4 activation. However, OPN also enhances Brk-dependent NIK/NF-nB–mediated ATF-4 activation. Transfection ofcells with wt or dn ATF-4 had no effect on OPN-induced NF-nB–DNA binding, indicating that OPN-regulated crosstalk betweenNF-nB and ATF-4 is unidirectional toward ATF-4 (SupplementaryFig. S1D). To determine the role of Brk and ATF-4 on OPN-inducedVEGF expression, cells were individually transfected with wt/KMBrk, Brk siRNA, or wt/dn ATF-4 and then treated with OPN.The results showed that cells transfected with wt Brk or wtATF-4 increased VEGF expression upon OPN treatment, whereastransfection with KM Brk, Brk siRNA or dn ATF-4 showed drasticsuppression of VEGF expression (Fig. 2D). Moreover, InBa sup. rep.significantly reduced OPN-induced VEGF expression (Fig. 2D). Theeffect of OPN on VEGF promoter activity was examined either bytransfecting the cells with VEGF-GFP reporter construct alone orcotransfecting with wt and KM Brk, wt and dn ATF-4, InBa sup.rep., or Brk siRNA and then treated with OPN. The OPN-inducedGFP expression was analyzed by fluorescence microscopy (Nikon),quantified by Image Pro Plus 6.0 Software (Nikon), and representedin the form of bar graph (Supplementary Fig. S2). These data clearlysuggested that Brk, ATF-4, and NF-nB play important role in OPN-induced VEGF expression both at transcriptional and protein level.

VEGF and its receptor NRP-1 play crucial roles in OPN-regulated breast tumor cell motility through autocrinemechanism. Recent data revealed that VEGF acts as survival,migratory and proliferative factor in breast carcinoma cells (10, 36).NRP-1 is the only VEGF-specific receptor that is expressed in MDA-MB-231 cells (29, 30). We therefore speculated that OPN mightregulate the interaction between VEGF and NRP-1, which maycontrol the motility of MDA-MB-231 cells. Accordingly, cells weretreated with OPN alone or along with anti-VEGF or anti-NRP-1blocking antibody, and lysates were immunoprecipitated with anti-VEGF antibody and immunoblotted with anti–NRP-1 antibody. Thedata suggested that OPN-induced VEGF interacts with NRP-1 inMDA-MB-231 cells through autocrine manner (Fig. 3A, I). Toexamine whether OPN has any effect on breast tumor cell motilityand whether VEGF and NRP-1 are involved in this process, woundassay was performed using MDA-MB-231 cells with typicalcobblestone morphology. Wounds with a constant diameter weremade, and cells were treated with OPN alone or along with anti-VEGF or anti–NRP-1 blocking antibody. In separate experiments,cells were transfected with NRP-1i and then treated with OPN, andwound assay was performed. The wound photographs were takenunder phase contrast microscope (Fig. 3A, II). Moreover, cellstransfected with wt Brk or wt ATF-4 enhanced, whereas KM Brk ordn ATF-4 drastically suppressed, OPN-induced wound migration

Figure 2. OPN regulates VEGF expression via Brk/NF-nB/ATF-4 signaling pathways in MDA-MB-231 cells. A, OPN induces Brk phosphorylation and kinaseactivity through avh3 integrin–dependent but c-Src and PI3K-independent pathway. Cells were treated with 0.5 Amol/L OPN for 8 min or pretreated with anti-avh3antibody or wortmannin or pp2 for 2 h and then treated with OPN. The cell lysate was immunoprecipitated with anti-Brk antibody. Half of the immunoprecipitatedsamples were analyzed by Western blot using antiphosphotyrosine antibody. Remaining half of the samples was used for detecting Brk kinase activity byautophosphorylation study. Total Brk expression in the cells was used as loading control. The levels of p-Tyr-Brk were quantified by densitometric analysis andrepresented in the form of bar graph (*, P < 0.03). B, OPN-induced Brk-mediated NIK phosphorylation was determined by transfecting the cells with wt or KM Brkand then treated with 0.5 Amol/L OPN for 15 min. Cell lysates were analyzed by Western blot using anti–phosphorylated NIK antibody. C, Brk regulates OPN-dependentATF-4–DNA binding and crosstalk between NF-nB and ATF-4. Cells were transfected with wt or KM Brk and treated with OPN. In separate experiments, cells weretransfected with wt NIK or mut NIK or sup. rep. InBa or wt Brk and mut NIK or KM Brk and wt NIK or wt Brk and InBa sup. rep followed by treatment with OPN.ATF-4–DNA binding activity was analyzed by EMSA. Note that Brk up-regulates ATF-4 activity through NF-nB–dependent/independent pathway. D, roles of Brkand ATF-4 in OPN-induced VEGF expression. Cells were individually transfected with wt and KM Brk or wt and dn ATF-4 or Brk siRNA or InBa sup. rep. and then treatedwith OPN. VEGF expressions (secreted and total) were analyzed by Western blot. All figures are representation of three experiments. Fold changes were calculated.



(Fig. 3A, II). Interestingly, the enhanced wound migration caused byoverexpression of wt Brk or wt ATF-4 was suppressed upon treat-ment of cells with anti-VEGF and anti–NRP-1 blocking antibodies(Fig. 3A, II). These results clearly indicated that OPN regulatesVEGF-dependent wound migration through NRP-1–mediatedautocrine mechanism.OPN-induced VEGF controls KDR-dependent endothelial

cell motility and in vivo angiogenesis through paracrine andjuxtacrine mechanisms. Tumor-derived VEGF interacts withendothelial cell surface VEGF receptor-2 (also called KDR) andinduces KDR phosphorylation that leads to tumor angiogenesis(37). Therefore, we sought to determine whether OPN-induced

VEGF, which is expressed in MDA-MB-231 cells, can regulate KDRphosphorylation in endothelial (EA.hy-926) cell line throughparacrine loop. Accordingly, MDA-MB-231 cells were treated withOPN and CM either alone or mixed with anti-VEGF blockingantibody incubated with endothelial cells. In separate experiments,CM of OPNi transfected MDA-MB-231 cells were incubated withendothelial cells. Cell lysates were analyzed by Western blot forKDR phosphorylation. The data revealed that OPN-induced VEGFenhanced KDR phosphorylation in endothelial cells throughparacrine mechanism (Supplementary Fig. S3A). To investigatewhether OPN-induced MDA-MB-231 cell-derived VEGF regulatesendothelial cell wound motility, MDA-MB-231 cells were either

Figure 3. A, OPN enhances the VEGF–NRP-1 interaction in MDA-MB-231 cells. I, OPN-induced VEGF–NRP-1 interactions were examined by treating thecells with OPN for 6 h or pretreated with anti-VEGF or anti–NRP-1 blocking antibody and then treated with OPN. Cell lysates were immunoprecipitated with anti-VEGFantibody and analyzed by Western blot with anti–NRP-1 antibody. II, VEGF and NRP-1 play crucial roles in OPN-induced tumor cell motility. Wounds with aconstant diameter were made, and cells were treated with OPN. In separate experiments, cells were transfected with wt and KM Brk or wt and dn ATF-4 and thentreated with OPN in absence or presence of anti-VEGF or anti–NRP-1 blocking antibody, and wound assay was performed. In separate experiments, cells wereindividually transfected with OPNi, and wound assay was conducted. Cells were also transfected with NRP-1i and then treated with OPN. Wound photographswere taken after termination of the experiments (t = 12 h). B, OPN controls VEGF-dependent endothelial cell (EC ) wound migration in a paracrine mechanism.Endothelial (EA-hy.926) cells were incubated with conditioned media obtained from OPN-treated or OPNi transfected MDA-MB-231 cells, and wound assay wasperformed and represented in the form of bar graph (*, P < 0.007; **, P < 0.01). In separate experiments, endothelial cells were also incubated with OPN-treatedconditioned media mixed with anti-VEGF blocking antibody and used for wound assay. C, OPN enhances VEGF-dependent NRP-1/pKDR mediated tumor-endothelialcell interaction through juxtacrine mechanism. Endothelial cells were transfected with NRP-1i (EA NRP-1i) and cocultured with normal or OPNi or NRP-1itransfected MDA-MB-231 cells. OPN alone or along with anti-VEGF blocking antibody were added to the cocultured cells. Cell lysates were immunoprecipitated withanti–NRP-1 antibody and analyzed by Western blot with anti–phosphorylated KDR antibody. D, OPN enhances VEGF/NRP-1–mediated endothelial cell migrationthrough juxtacrine manner. Endothelial cells were added on the upper portion of modified Boyden chamber. MDA-MB-231 cells seeded in lower chamber were eithertreated with OPN alone or along with anti-VEGF or anti–NRP-1 blocking antibody and used for direct comigration assays. In separate experiments, MDA-MB-231 cellswere transfected with OPNi or NRP-1i and seeded in the lower chamber, and comigration assays were performed. Endothelial cells were migrated from upperchamber to the lower chamber and counted under inverted microscope (Nikon). Columns, means of three determinations; bars, SE (*, P < 0.001; **, P < 0.004).

Cancer Research


treated or transfected under similar conditions as described aboveand endothelial cell wound migration was performed. The data areanalyzed and represented in the form of bar graph (Fig. 3B). Theseresults showed that OPN enhances VEGF-dependent endothelialcell motility in paracrine mechanism.The interaction of OPN-induced VEGF with NRP-1 and KDR

through juxtacrine mechanism is shown by coculture experimentsusing MDA-MB-231 and endothelial cells. Accordingly, endothelialcells were transfected with NRP-1i and cocultured with MDA-MB-231 cells in the presence of OPN alone or along with anti-VEGFantibody. In separate experiments, MDA-MB-231 cells were alsotransfected with OPNi or NRP-1i and used for coculture experi-ments. The cell lysates were immunoprecipitated with anti–NRP-1antibody, followed by immunoblotting with anti–pKDR antibody.Our data revealed that OPN-induced VEGF acts as bridge betweenNRP-1, expressed in tumor cells, and KDR, expressed in endothelialcells, and regulates NRP-1–mediated KDR phosphorylation throughjuxtacrine mechanism (Fig. 3C). To verify whether pKDR is indeedexpressed in endothelial but not in MDA-MB-231 cells, endothelialcells were incubated with CM of OPN-treated MDA-MB-231 cells.In separate experiments, MDA-MB-231 cells were also treated with

OPN. The level of pKDR was analyzed by Western blot. The resultsshowed that pKDR/KDR is detectable only in endothelial but not inMDA-MB-231 cells (Supplementary Fig. S3B). The endothelial-breast tumor cell interaction was further examined by directcomigration assay. The MDA-MB-231 cells either treated withexogenous OPN or transfected with OPNi or NRP-1i or treated withanti-VEGF or anti–NRP-1 antibody followed by treatment with OPNand used in the lower chamber whereas endothelial cells were usedin the upper portion of modified Boyden chamber. The resultsshowed that OPN stimulates VEGF/NRP-1–mediated endothelialcell migration through juxtacrine pathway (Fig. 3D). Taken toge-ther, these data showed that OPN-induced VEGF interacts withendothelial cell surface receptor, KDR, and regulates endothelialcell motility through paracrine and juxtacrine mechanisms.OPN promotes VEGF-dependent in vivo angiogenesis. To

assess whether OPN regulates VEGF-dependent in vivo angiogen-esis, Matrigel plug assay was performed. The CM obtained fromOPN-treated or OPNi-transfected MDA-MB-231 cells was mixedwith Matrigel either alone or along with anti-VEGF blockingantibody and injected s.c. (ventral groin region) to the nude mice.Matrigel plugs were excised, and the sections were analyzed by

Figure 4. OPN induces VEGF/NRP-1–dependent tumor growth and angiogenesis in nude mice model. A, OPN regulates VEGF-dependent angiogenesis inMatrigel plug assay. The OPNi-transfected or OPN-treated CM of MDA-MB-231 cells either alone or along with anti-VEGF antibody were mixed with Matrigel and theninjected s.c. to ventral groin region of nude mice (NMRI, nu/nu). Matrigel plugs were excised, and the sections were subjected to histopathologic analysis. Notethat endothelial cell nuclei stained blue were observed inside the bed of pink stained Matrigel. The endothelial cells were further characterized by immunofluorescenceusing anti-vWF antibody (stained with Cy3; red) as marker and visualized under confocal microscopy. The phosphorylation of KDR was also detected byimmunofluoresence by using anti–phosphorylated KDR antibody followed by staining with FITC (green ). Nuclei were stained with 4¶,6-diamidino-2-phenylindole(DAPI; blue ). B, typical photographs of orthotopic breast tumors in athymic nude mice model. Bottom, excised tumors of respective mice. Six mice were usedin each set of experiments. C, the levels of VEGF, OPN, and CA 15-3 from xenograft tumor lysates and sera were analyzed by Western blot. NF-nB andATF-4–DNA-binding in these orthotopic tumors were analyzed by EMSA. Actin was used as control. D, mice breast tumors were analyzed by histopathology (top )and immunohistochemistry using anti-vWF antibody (bottom ). vWF was stained with Cy3 (red) and nuclei were counterstained with DAPI (blue ). Normal mammary fatpad (MFP ) is used as control.



histopathologic and immunohistochemical studies using anti-vWFand anti-pKDR antibodies. The data revealed that OPN enhancesVEGF-dependent in vivo angiogenesis through KDR phosphoryla-tion (Fig. 4A). The neovascularization and KDR phosphorylationwere quantified and represented graphically (SupplementaryFig. S4A). The data showed that OPN regulates VEGF-dependentin vivo angiogenesis through paracrine loop, whereas silencingtumor-derived OPN resulted in drastic reduction of tumorangiogenesis.VEGF plays a crucial role in OPN-induced tumor growth and

angiogenesis in orthotopic mice model. Our in vitro and in vivodata prompted us to examine whether OPN regulates tumorgrowth in orthotopic nude mice model and whether VEGF isinvolved in this process. Accordingly, MDA-MB-231 cells weretransfected with OPN cDNA or treated with OPN and injectedorthotopically into the mammary fat pad of female nude mice. Inseparate experiments, OPNi or NRP-1i were mixed with transfec-tion reagent and injected to the site of the tumor. Purified OPN andanti-VEGF blocking antibody were injected intratumorally twice aweek up to completion of the experiments. After 6 weeks, micewere sacrificed and tumors were collected. The typical photographsof tumors are shown in Fig. 4B . The changes of tumor weight

(% to control) were analyzed and represented as bar graph(Supplementary Fig. S4B). Blocking of VEGF with its antibody orNRP-1i significantly suppressed OPN-induced breast tumor growthin nude mice suggesting that VEGF and its receptor NRP-1 playcrucial role in this process.Tumor sections were also analyzed by histopathology. The

results showed that tumors generated by OPN exhibit enhancedtumor cell infiltration, poor differentiation, higher nuclearpolymorphism, mitotic count (12–16/hpf) and increased vesselformation in mammary fat pad of mice compared with control(Fig. 4D, top). Moreover, well-differentiated tumor structure, littleinfiltration and nuclear polymorphism, lesser mitotic features(2-5/hpf), and decreased vessel formation were observed in thetumors of the mice injected with anti-VEGF blocking antibody orNRP-1i along with OPN. To examine whether OPN regulates tumorangiogenesis, mice tumors were analyzed by immunohistochemis-try using anti-vWF antibody (Fig. 4D, bottom ; SupplementaryFig. S4C). The results indicated that OPN enhanced microvesseldensity, whereas blocking of VEGF or silencing NRP-1 suppressedOPN-induced tumor angiogenesis. Silencing tumor-derived OPNdrastically suppressed vWF expression, suggesting the potentialrole of tumor-derived OPN in tumor angiogenesis. To examine

Figure 5. A and B, analysis of human breast tumor clinical specimens with different pathologic grades. Human breast tumor specimens (1–14) were collectedfrom local hospital with informed consent. The tumor grading was performed by histopathologic analysis using a modified Scarff-Bloom-Richardson system. Tumorsamples were analyzed by Western blot using anti-VEGF or anti-OPN antibody (A). The levels of OPN and VEGF were also detected in normal breast epithelialtissues. The expressions of OPN and VEGF were quantified and represented in the form of bar graph (B ). C and D, histopathologic micrographs of normal andbreast tumor tissues of different grades (specimen 5, grade II; specimens 8 and 9, grade III). The sections of normal and tumor tissues (specimens 5, 8, and 9)of different grades were analyzed by immunohistochemistry using anti-vWF antibody. Note that the enhanced microvessel density was observed in higher gradesof tumor (specimens 8 and 9) and correlates with increased OPN and VEGF expressions. The vWF-positive areas indicate the microvessel density (pink arrows ).Cellular localization of NF-nB and expression profiles of NRP-1, ATF-4, and Brk were analyzed by immunohistochemical studies using their specific antibodies.Note that NF-nB, ATF-4, and Brk are stained with FITC (green ), whereas NRP-1 is stained with TRITC (red).

Cancer Research


whether OPN enhances the interaction between VEGF and NRP-1in vivo system, the tumor lysates were immunoprecipitated withanti-VEGF antibody and analyzed by Western blot using anti–NRP-1 antibody. Treatment with anti-VEGF blocking antibody orsilencing NRP-1 significantly suppressed OPN-induced interactionbetween VEGF and NRP-1 (data not shown) and further suggestedthat OPN augments the interaction between VEGF and NRP-1 thatultimately control the tumor growth and angiogenesis.Tumor tissues were lysed, and the levels of OPN and VEGF were

detected by Western blot (Fig. 4C). The level of CA15-3 from miceserum was analyzed as breast cancer marker. The levels of VEGFand OPN were also detected from mice serum (Fig. 4C). Significantinduction of VEGF was observed in tumor generated by OPN.Tumor samples were also analyzed by EMSA (Fig. 4C) and immu-nofluorescence (Supplementary Fig. S4D), and the data suggestedthat OPN augments NF-nB and ATF-4–DNA binding and nuclearlocalization in these tumors. Moreover, silencing tumor-derivedOPN significantly suppressed NF-nB and ATF-4–DNA binding andnuclear localization. These data suggested that both tumor-derivedand exogeneous OPN regulates NF-nB and ATF-4–dependent VEGF/NRP-1–mediated breast tumor progression and angiogenesis.Expressions of OPN and VEGF and their implications in

human breast tumor angiogenesis. The in vitro and in vivo datafurther prompted us to extend these studies in human clinicalbreast tumor specimens. Human solid breast tumor specimens andregional lymph nodes were collected with informed consent from alocal hospital. Tumor grading was performed with the help of

expert oncopathologist using modified Scarff-Bloom-Richardsonsystem, and regional lymph node metastasis was analyzed byhistopathology (Supplementary Table S1). The expressions of OPNand VEGF (Fig. 5A) and the interaction between VEGF and NRP-1(data not shown) were studied in tumor specimens of differentgrades. The levels of OPN and VEGF were quantified and repre-sented as bar graph (Fig. 5B). The data revealed that there weresignificantly higher levels of OPN and VEGF expressions (Fig. 5A ;8 of 14 tumor specimens and one normal sample), and these datacorrelated with tumors of higher grades. The data also indicatedthat the interaction between VEGF and NRP-1 was also enhancedsignificantly with tumors of higher grades (data not shown). Theimmunohistochemical staining with anti-vWF antibody showed theenhanced angiogenic vessel formation in breast tumor specimensof higher grades (Fig. 5C).Localization and expression of NF-KB, ATF-4, Brk, and NRP-1

in breast tumors specimens. To examine the activation statusof NF-nB in different grades of tumor specimens, tissue sectionswere analyzed by immunohistochemistry using anti–NF-nB, p65antibody. The results indicated that the NF-nB nuclear localizationwas significantly enhanced in higher grades (grades II and III) oftumors compared with the tumor of lower grades or normal breasttissues (Fig. 5C). The status of expressions of NRP-1, ATF-4, andBrk in different grades of tumors was also analyzed by immu-nohistochemistry using their specific antibodies. These resultsindicated that the higher expressions of NRP-1, ATF-4, and Brkcorrelate with higher grades of tumors (Fig. 5D). The ATF-4 nuclear

Figure 6. A schematic representation of OPN-induced NF-nB/ATF-4–mediated VEGF expression through activation of Brk/NIK leading to tumor growth andangiogenesis through autocrine, paracrine, and juxtacrine mechanisms.



localization was also significantly enhanced in higher grades(grades II and III) of tumors and correlated with NF-nB (Fig. 5Cand D). The status of OPN and VEGF expressions and its corre-lation with NF-nB and ATF-4 activation and metastatic potentialin various grades of tumors are summarized in SupplementaryTable S1. These results showed that tumors with higher gradesexhibit significantly higher levels of ATF-4 and NF-nB nuclearlocalization, and these data correlated with enhanced OPN, Brk,and VEGF expression both in in vitro and in vivo systems, whichfurther corroborated with tumor growth and angiogenesis in breastcarcinoma.

Discussion

We have for the first time shown the molecular mechanism ofOPN-induced VEGF expression and its potential role in regulatingin vitro cell motility which ultimately modulates in vivo tumorgrowth and angiogenesis in breast cancer model. Our data alsorevealed that OPN might be necessary for maximal induction ofneovascularization by inducing VEGF expression through activa-tion of Brk/NF-nB/ATF-4 pathways. Several lines of evidencessupport this conclusion. First, OPN regulates VEGF promoteractivity and expression in human breast carcinoma cell lines.Second, OPN stimulates Brk phosphorylation and Brk/NIK-mediated NF-nB–dependent/independent ATF-4 activation andcrosstalk between NF-nB and ATF-4 that leads to VEGF expression.Third, in autocrine pathway, OPN-induced VEGF interacts withtumor cell surface receptor NRP-1 and stimulates VEGF–NRP-1–mediated tumor cell motility. Fourth, in paracrine loop, OPN-induced VEGF up-regulates KDR phosphorylation, resulting inenhanced endothelial cell motility and in vivo angiogenesis in nudemice. Fifth, in juxtacrine manner, OPN-induced VEGF regulatestumor-endothelial cell interaction through binding with NRP-1 inbreast cancer and KDR in endothelial cells. Sixth, OPN-inducedVEGF leads to increased breast tumor growth and angiogenesis innude mice model and blocking of tumor-derived VEGF or silencingof tumor-derived OPN and NRP-1 significantly suppressed breasttumor progression and angiogenesis. Seventh, analysis of humansolid breast tumor specimens showed the significant correlationbetween the OPN, Brk, NRP-1, and VEGF expressions NF-nB andATF-4 activations with different pathologic grades of tumors. Allthese data supported by the fact that OPN and its downstreamsignaling molecules might be used as therapeutic target for thetreatment of breast cancer.Previous reports indicated that Brk plays important role during

breast tumor progression (22, 23). We provide evidence, at least inpart, that Brk plays crucial role in OPN-induced NIK-mediatedNF-nB activation. Our data also showed that OPN regulates Brk-dependent ATF-4 activation and a crosstalk between NF-nB andATF-4 that controls VEGF-dependent tumor angiogenesis. NRP-1acts as key receptor for VEGF-mediated signaling in MDA-MB-231cells (36, 38). Recent evidence indicated that NRP-1 could inde-pendently regulate VEGF-induced cell signaling (39). We sought todetermine whether OPN plays any role in regulating the interactionbetween VEGF and NRP-1 in breast cancer cells. Our resultsshowed that enhanced interaction of VEGF and NRP-1 in responseto OPN control the breast cancer cell migration and tumor growththrough autocrine mechanism. We have also shown that OPN-induced tumor-derived VEGF regulates KDR/NRP-1–mediatedendothelial cell motility and tumor angiogenesis through paracrineand juxtracrine mechanisms.

The role of OPN in the regulation of tumor angiogenesis is underintense investigation (12, 40). Earlier Senger et al. reported thatVEGF induces OPN expression in endothelial cells, which in turnregulates angiogenesis (41). However we have shown that OPNregulates VEGF expression, which controls tumor angiogenesisthrough autocrine, paracrine, and juxtacrine pathways. These datasuggested that there could be a positive feedback loop betweenOPN and VEGF which might play an important role in tumorprogression and angiogenesis. It has also been reported thatcoexpression of VEGF and OPN correlates with angiogenesis inpatients with stage I lung adenocarcinoma (42). Earlier reportshave shown that cooperation of OPN and VEGF may facilitatetumor progression and angiogenesis in carcinomas derived fromother organs (43). Recently, we have showed that OPN regulatescyclooxygenase-2–dependent prostaglandin E2 production, whichin turn regulates prostate cancer angiogenesis (14). Althoughearlier reports indicated that OPN acts as an angiogenic factor butthe molecular mechanism by which OPN regulates VEGF-dependent tumor angiogenesis in breast cancer is not well defined.Our study for the first time showed the in-depth signaling pathwaysby which OPN induces cell migration and angiogenesis throughinduction of VEGF expression in breast cancer. Moreover, thesestudies correlate with the data obtained from human breast tumorspecimens in Indian scenario.Suppression of tumor progression by targeting the angiogenic

switch is one of the greatest challenges for cancer biologists andmedical oncologists. Recent reports on anticancer therapyindicated that suppression of tumor angiogenesis by administra-tion of antiangiogenic molecules is one of the most potenttherapeutic approaches for the treatment of cancer (44). However,it is also known that some side effect of these antiangiogenic drugsare considered as the main problem of this therapy (45). Faced withthe barrier to the traditional antiangiogenic strategies, we tested anin vivo RNA interference–based approach by targeting OPN andNRP-1. In recent time, in vivo siRNA-based approach has beenproposed for treatment of several diseases including cancer (46).Therefore, appropriately designed OPN siRNA, injected intra-tumorally, exhibit significant suppression of tumor angiogenesisand may help in the surgical excision of tumor and provideconsiderable therapeutic value for the treatment of cancer.Finally, we have shown for the first time that OPN induced the

potent angiogenic molecule, VEGF expression. OPN regulatestumor growth and angiogenesis by induction of VEGF expressionthrough activation of Brk/NF-nB/ATF-4 pathways. These datashowed that OPN-induced VEGF regulates tumor angiogenesisthrough autocrine, paracrine, and juxtacrine mechanisms (Fig. 6).Our results further warrant that the mechanism shown in themouse model underlies the human pathology, and a clear under-standing of such mechanism(s) may facilitate the development ofnovel therapeutic approaches to suppress OPN-regulated NF-nB/ATF-4–dependent VEGF expression, thereby controlling tumorgrowth and angiogenesis.

Acknowledgments

Received 6/12/2007; revised 10/5/2007; accepted 10/29/2007.Grant support: Department of Biotechnology, Government of India (G.C. Kundu)

and Council of Scientific and Industrial Research, Government of India (G.Chakraborty and S. Jain).

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

Cancer Research




References

1. Folkman J, Shing Y. Angiogenesis. J Biol Chem 1992;267:10931–4.2. Rifkin DB, Moscatelli D. Recent developments in thecell biology of basic fibroblast growth factor. J Cell Biol1989;109:1–6.3. Nicosia RF, Nicosia SV, Smith M. Vascular endothelialgrowth factor, platelet-derived growth factor, andinsulin-like growth factor-1 promote rat aortic angio-genesis in vitro . Am J Pathol 1994;145:1023–9.4. Gleave ME, Hsieh JT, Wu HC, et al. Epidermal growthfactor receptor-mediated autocrine and paracrine stim-ulation of human transitional cell carcinoma. CancerRes 1993;53:5300–7.5. McMahon G. VEGF receptor signaling in tumorangiogenesis. Oncologist 2000;5:3–10.6. Ferrara N. VEGF and the quest for tumor angiogenesisfactors. Nat Rev Cancer 2002;2:795–803.7. Ferrara N, Gerber HP, LeCouter J. The biology of VEGFand its receptors. Nat Med 2003;9:669–76.8. Cross MJ, Dixelius J, Matsumoto T, Claesson-Welsh L.VEGF-receptor signal transduction. Trends Biochem Sci2003;28:488–94.9. Banerjee SN, Sengupta K, Banerjee S, Saxena NK,Banerjee SK. 2-Methoxyestradiol exhibits a biphasiceffect on VEGF-A in tumor cells and upregulation ismediated through ER-a, a possible signaling pathwayassociated with the impact of 2-ME2 on proliferativecells. Neoplasia 2003;5:417–26.10. Mercurio AM, Bachelder RE, Bates RC, Chung J.Autocrine signaling in carcinoma: VEGF and the a 6 h 4integrin. Semin Cancer Biol 2004;14:115–22.11. Sodek J, Ganss B, McKee MD. Osteopontin. Crit RevOral Biol Med 2000;11:279–303.12. Chakraborty G, Jain S, Behera R, et al. Themultifaceted roles of osteopontin in cell signaling,tumor progression and angiogenesis. Curr Mol Med2006;6:819–30.13. Wai PY, Kuo PC. The role of osteopontin in tumormetastasis. J Surg Res 2004;121:228–41.14. Jain S, Chakraborty G, Kundu GC. The crucial role ofcyclooxygenase-2 in osteopontin-induced protein kinaseC a/c-Src/InB kinase a/h-dependent prostate tumorprogression and angiogenesis. Cancer Res 2006;66:6638–48.15. Philip S, Bulbule A, Kundu GC. Osteopontin stim-ulates tumor growth and activation of promatrixmetalloproteinase-2 through nuclear factor-n B-mediat-ed induction of membrane type 1 matrix metallopro-teinase in murine melanoma cells. J Biol Chem 2001;276:44926–35.16. Das R, Mahabeleshwar GH, Kundu GC. Osteopontinstimulates cell motility and nuclear factor nB-mediatedsecretion of urokinase type plasminogen activatorthrough phosphatidylinositol 3-kinase/Akt signalingpathways in breast cancer cells. J Biol Chem 2003;278:28593–606.17. Rangaswami H, Bulbule A, Kundu GC. Nuclear

factor-inducing kinase plays a crucial role in osteopon-tin-induced MAPK/I n B a kinase-dependent nuclearfactor n B-mediated promatrix metalloproteinase-9activation. J Biol Chem 2004;279:38921–35.18. Rangaswami H, Bulbule A, Kundu GC. Osteopontin:role in cell signaling and cancer progression. Trends CellBiol 2006;16:79–87.19. Rittling SR, Chen Y, Feng F, Wu Y. Tumor-derivedosteopontin is soluble, not matrix associated. J BiolChem 2002;277:9175–82.20. Senger DR, Perruzzi CA, Papadopoulos A, Tenen DG.Purification of a human milk protein closely similar totumor-secreted phosphoproteins and osteopontin. Bio-chim Biophys Acta 1989;996:43–8.21. Denhardt DT, Mistretta D, Chambers AF, et al.Transcriptional regulation of osteopontin and themetastatic phenotype: evidence for a Ras-activatedenhancer in the human OPN promoter. Clin ExpMetastasis 2003;20:77–84.22. Zhang P, Ostrander JH, Faivre EJ, Olsen A, Fitzsim-mons D, Lange CA. Regulated association of proteinkinase B/Akt with breast tumor kinase. J Biol Chem2005;280:1982–91.23. Chen HY, Shen CH, Tsai YT, Lin FC, Huang YP, ChenRH. Brk activates rac1 and promotes cell migration andinvasion by phosphorylating paxillin. Mol Cell Biol 2004;24:10558–72.24. Harvey AJ, Crompton MR. Use of RNA interference tovalidate Brk as a novel therapeutic target in breastcancer: Brk promotes breast carcinoma cell prolifera-tion. Oncogene 2003;22:5006–10.25. Roybal CN, Yang S, Sun CW, et al. Homocysteineincreases the expression of vascular endothelial growthfactor by a mechanism involving endoplasmic reticulumstress and transcription factor ATF4. J Biol Chem 2004;279:14844–52.26. Roybal CN, Hunsaker LA, Barbash O, Vander Jagt DL,Abcouwer SF. The oxidative stressor arsenite activatesvascular endothelial growth factor mRNA transcriptionby an ATF4-dependent mechanism. J Biol Chem 2005;280:20331–9.27. He Z, Tessier-Lavigne M. Neuropilin is a receptor forthe axonal chemorepellent semaphorin III. Cell 1997;90:739–51.28. Kolodkin AL, Levengood DV, Rowe EG, Tai YT, GigerRJ, Ginty DD. Neuropilin is a semaphorin III receptor.Cell 1997;90:753–62.29. Soker S, Fidder H, Neufeld G, Klagsbrun M.Characterization of novel vascular endothelial growthfactor (VEGF) receptors on tumor cells that bindVEGF165 via its exon 7-encoded domain. J Biol Chem1996;271:5761–7.30. Soker S, Takashima S, Miao HQ, Neufeld G, KlagsbrunM. Neuropilin-1 is expressed by endothelial and tumorcells as an isoform-specific receptor for vascularendothelial growth factor. Cell 1998;92:735–45.31. Miao HQ, Lee P, Lin H, Soker S, Klagsbrun M.Neuropilin-1 expression by tumor cells promotes tumorangiogenesis and progression. FASEB J 2000;14:2532–9.

32. Barr MP, Byrne AM, Duffy AM, et al. A peptidecorresponding to the neuropilin-1-binding site on VEGF(165) induces apoptosis of neuropilin-1-expressingbreast tumour cells. Br J Cancer 2005;92:328–33.33. Leopold JA, Walker J, Scribner AW, et al. Glucose-6-phosphate dehydrogenase modulates vascular endothe-lial growth factor-mediated angiogenesis. J Biol Chem2003;278:32100–6.34. Sweeney CJ, Mehrotra S, Sadaria MR, et al. Thesesquiterpene lactone parthenolide in combination withdocetaxel reduces metastasis and improves survival in axenograft model of breast cancer. Mol Cancer Ther2005;4:1004–12.35. Chakraborty G, Rangaswami H, Jain S, Kundu GC.Hypoxia regulates cross-talk between Syk and Lckleading to breast cancer progression and angiogenesis.J Biol Chem 2006;281:11322–31.36. Bachelder RE, Crago A, Chung J, et al. Vascularendothelial growth factor is an autocrine survival factorfor neuropilin-expressing breast carcinoma cells. CancerRes 2001;61:5736–40.37. Matsumoto T, Claesson-Welsh L. VEGF receptorsignal transduction. Sci STKE 2001;RE 21:1–17.38. Bachelder RE, Lipscomb EA, Lin X, et al. Competingautocrine pathways involving alternative neuropilin-1ligands regulate chemotaxis of carcinoma cells. CancerRes 2003;63:5230–3.39. Wang L, Zeng H, Wang P, Soker S, Mukhopadhyay D.Neuropilin-1-mediated vascular permeability factor/vascular endothelial growth factor-dependent endothe-lial cell migration. J Biol Chem 2003;278:48848–60.40. Takahashi F, Akutagawa S, Fukumoto H, et al.Osteopontin induces angiogenesis of murine neuroblas-toma cells in mice. Int J Cancer 2002;98:707–12.41. Senger DR, Ledbetter SR, Claffey KP, Papadopoulos-Sergiou A, Peruzzi CA, Detmar M. Stimulation ofendothelial cell migration by vascular permeabilityfactor/vascular endothelial growth factor through co-operative mechanisms involving the avh3 integrin,osteopontin, and thrombin. Am J Pathol 1996;149:293–305.42. Shijubo N, Uede T, Kon S, et al. Vascular endo-thelial growth factor and osteopontin in stage I lungadenocarcinoma. Am J Respir Crit Care Med 1999;160:1269–73.43. Shijubo N, Uede T, Kon S, Nagata M, Abe S. Vascularendothelial growth factor and osteopontin in tumorbiology. Crit Rev Oncog 2000;11:135–46.44. Giavazzi R, Nicoletti MI. Small molecules in anti-angiogenic therapy. Curr Opin Invest Drugs 2002;3:482–91.45. Kerbel RS, Viloria-Petit A, Klement G, Rak J.‘Accidental’ anti-angiogenic drugs. anti-oncogene direct-ed signal transduction inhibitors and conventionalchemotherapeutic agents as examples. Eur J Cancer2000;36:1248–57.46. Jain S, Chakraborty G, Bulbule A, Kaur R, Kundu GC.Osteopontin: an emerging therapeutic target for anti-cancer therapy. Expert Opin Ther Targets 2007;11:81–90.

Abstract. The development and progression of malignanttumors depends on the formation of new blood vessels insidethe tumor. This phenomenon is termed tumor angiogenesis.Angiogenesis is one of the fundamental processes that occurduring cancer progression, and depends on the expressionand activation of various angiogenic molecules, cytokines,growth factors, kinases and transcription factors. We recentlydemonstrated that the chemokine-like ECM-associatedprotein osteopontin (OPN) turns on the angiogenic switch byupregulating expression of vascular endothelial growth factor(VEGF) in a human breast cancer model. Furthermore, weproposed that targeting OPN-induced VEGF expression couldbe a potential therapeutic approach for the treatment of breastcancer. In this study, we demonstrate that curcumin (diferuloyl-methane) abrogates OPN-induced VEGF expression and curbsOPN-induced VEGF-dependent breast tumor angiogenesisin vivo. We also explore the fact that curcumin in combi-nation with anti-VEGF or anti-neuropilin (NRP)-1 antibodyexhibits enhanced anti-angiogenic activity compared to cur-cumin alone. Our results indicate that curcumin suppressesOPN-induced VEGF expression and tumor angiogenesis, andsuggest that this study may aid in the development of acurcumin-based OPN-targeted therapeutic approach to thecontrol of breast tumor angiogenesis.

Introduction

Breast cancer is considered to be one of the most commoncancer threats worldwide. According to the World HealthOrganization (WHO), it is the most deadly cancer in females,with more than 500,000 deaths attributed to it globally eachyear. To date, the major cause of breast cancer is not clearlyunderstood. However, experimental evidence has revealedthat the expression profiles of certain oncogenic molecules or

biomarkers are significantly associated with breast cancerprogression, and have shown great promise in breast cancertherapy (1,2). Osteopontin (OPN), a chemokine-like ECM-associated small integrin binding ligand N-linked glycoprotein(SIBLING), has recently been identified as one of the majormarkers of breast cancer progression (3,4). Elevated expres-sion of OPN at tumor sites as well as in the serum of breastcancer patients has signalled the prognostic importance ofthis protein in breast cancer (5,6). Moreover, the targeting ofOPN and its downstream signaling pathways has shown greatpromise in the therapeutics of various cancers, breast cancerincluded (7). The role of OPN in breast cancer progression istherefore being intensely investigated.

Angiogenesis, or the formation of new blood vessels fromexisting ones, is a key step in tumor growth, survival, progres-sion and metastasis (8). Tumor angiogenesis is thought toresult from the secretion of ‘angiogenetic factors’ by tumorcells. These include growth factors, cytokines and a number ofsmall molecules (9,10). Of these, vascular endothelial growthfactor (VEGF) has been recognized as the most important(11). Highly malignant tumors are characterized by enhancedvascularization, which is further correlated with elevatedVEGF expression (12). Previous reports suggest that tumor-derived VEGF interacts with tumor or endothelial cell surfacereceptors via autocrine or paracrine mechanisms and promotestumor angiogenesis (13-15). We recently demonstrated thatOPN augments VEGF expression and promotes VEGF-dependent breast tumor angiogenesis (16). Therefore, we canhypothesize that targeting OPN-induced VEGF might serveas a potential therapeutic approach for the treatment of breastcancer.

Curcumin (diferuloylmethane) is a polyphenol derivedfrom the rhizomes of Curcuma Longa, traditionally used as ananti-inflammatory compound, and appears to be useful in theprevention and treatment of various cancers (17-19). Curcuminhas exhibited a significant inhibitory effect on several malig-nant cancers, including breast cancer (20). It has also beenreported that curcumin downregulates the activation of NF-κBand suppresses tumor growth in various cancer models (21-23).We have previously reported that curcumin suppressed OPN-induced MMP-2 activation and tumor growth in a murinemelanoma model (24). However, the role of curcumin in theregulation of OPN-induced VEGF-dependent breast tumorangiogenesis is not well defined. In this study, we demonstratethat curcumin abrogates OPN-induced VEGF expression andsuppresses VEGF-dependent breast tumor angiogenesis.

MOLECULAR MEDICINE REPORTS 1: 641-646, 2008 641

Curcumin suppresses breast tumor angiogenesis byabrogating osteopontin-induced VEGF expression

GOUTAM CHAKRABORTY, SHALINI JAIN, SMITA KALE, REMYA RAJA,SANTOSH KUMAR, ROSALIN MISHRA and GOPAL C. KUNDU


Received April 9, 2008; Accepted June 2, 2008

DOI: 10.3892/mmr_00000005

_________________________________________

Correspondence to: Dr Gopal C. Kundu, National Center for CellScience, Pune 411 007, IndiaE-mail: [email protected]

Key words: osteopontin, vascular endothelial growth factor,curcumin, human breast cancer, tumor angiogenesis

641-646 29/7/08 11:39 Page 641

Materials and methods

Cell culture. The human breast adenocarcinoma cell lineMDA-MB-231 was obtained from American Type CultureCollection (Manassas, VA, USA). Cells were cultured in L-15media supplemented with 10% FBS, 100 μg/ml streptomycinand 100 units/ml penicillin. Human umbilical vein endothelialcells (HUVEC) were obtained from Lonza (Walkersville, MD,USA) and cultured in the EBM bullet kit (Lonza) accordingto the manufacturer's instructions.

Western blot analysis and EMSA. Western blot analysis andEMSA were performed as previously described (16,25).Briefly, the cells were treated with 0.5 μM OPN, or pretreatedwith various doses of curcumin (0-50 μM) and then treatedwith OPN. Cell lysates were used to detect VEGF expressionby Western blot analysis, and the nuclear extracts were usedfor NF-κB and ATF-4-DNA binding by EMSA.

Immunofluorescence study. Immunofluorescence staining wasperformed using a standard protocol as described previously(26). Briefly, the cells were treated with 0.5 μM OPN, orpretreated with 20 μM curcumin and then treated with OPN.Cells were fixed in 2% paraformaldehyde by incubation at4˚C for 15 min. After quenching with 0.1% glycine, cellswere permeabilized with 0.1% Triton X-100, washed threetimes with PBS for 5 min and then incubated in 2% BSA inPBS (pH 7.4) for 1 h at room temperature (RT) to block non-specific binding. Subsequently, cells were incubated withspecific primary antibody in 0.2% BSA for 2 h at RT, followedby incubation with secondary antibody for 1 h at RT. Nucleiwere stained with PI. Cells were washed and mounted inmounting media and analyzed using a confocal microscope(Zeiss).

Wound migration assay. The motility of MDA-MB-231 cellswas determined by wound migration assay as previouslydescribed (16,27). Cells were treated with 0.5 μM OPN, orpretreated with 0-20 μM curcumin followed by treatment withOPN. In separate experiments, cells were pretreated with10 μM curcumin along with anti-VEGF or anti-neuropilin(NRP)-1 antibody and then treated with OPN. After 12 h,wound photographs were captured using a phase contrastmicroscope (Nikon).

Co-migration assay. Tumor-endothelial cell interactions werecarried out using a co-migration assay with a modified Boydenchamber. MDA-MB-231 cells were treated with 0.5 μMOPN alone and used in the lower chamber. Endothelial cells(HUVEC) were seeded on the upper chamber. In separateexperiments, MDA-MB-231 cells were pretreated withcurcumin and then treated with OPN alone or with anti-VEGF or anti-NRP-1 antibody in the lower chamber, andco-migration experiments were performed. The migratedendothelial cells were stained with Giemsa, photographedand then counted in three high-power fields under an invertedmicroscope (Nikon). They were then analyzed statistically, andthe resulting data was represented by a bar graph.

In vivo Matrigel-based angiogenesis assay. The in vivoMatrigel angiogenesis assay was performed as previously

described (16). Briefly, Matrigel (0.5 ml) was subcutaneouslyinjected into the ventral groin region of female athymic NMRI(nu/nu) mice. In separate experiments, the conditioned medium(CM) of untreated or OPN-treated MDA-MB-231 cells wasmixed with Matrigel and injected into the mice. In other exper-iments, OPN-treated CM was mixed with curcumin (20 μM)in Matrigel and then injected into the mice. After 3 weeks,Matrigel plugs were excised and processed for histopathologyand immunohistochemistry using specific antibodies.

Immunohistochemical study. Immunohistochemical analysisof paraffin-embedded sections was performed using standardprocedure as previously described (16,27). After deparaf-finization in xylene, the sections were re-hydrated usingdescending grades of ethanol (100, 95 and 70%), washedthrice in PBS, then used for antigen retrieval. After quenchingwith 100 mM glycine, tissue sections were permeabilized with0.1% Triton X-100, washed twice with PBS, then blocked with2% bovine serum albumin for 1 h. Finally, the sections wereincubated with specific primary antibodies overnight at 4˚C.Subsequently, the slides were washed in PBS and incubatedwith specific secondary antibodies for 2 h. The sections weremounted and analyzed using a confocal microscope (Zeiss).

Statistical analysis. Bands were analyzed densitometrically(Kodak Digital Science) and fold changes were calculated.The wound migration and co-migration assays were analyzedstatistically, and the data were represented by means of a bargraph. Statistical differences were determined by the pairedStudent's t-test. Differences were considered significant atP-values <0.05.

Results

Curcumin abrogates OPN-induced VEGF expression. Wepreviously reported that curcumin suppressed OPN-inducedMMP-2 activation in a murine melanoma model (24). Morerecently, we demonstrated that OPN augments VEGF expres-sion in human breast cancer cells (16). To examine the effect ofcurcumin on OPN-induced VEGF expression, serum-starvedMDA-MB-231 cells were either treated with 0.5 μM OPN orwere pretreated with curcumin and then treated with OPN. Theresults indicate that curcumin significantly abrogates OPN-induced VEGF expression in these cells, with maximuminhibition observed using 20 μM of curcumin (Fig. 1A).

Effects of curcumin on OPN-induced NF-κB and ATF-4-DNAbinding. We recently showed that OPN promotes VEGFexpression via activation of NF-κB in MDA-MB-231 cells(16). However, it has been reported that curcumin suppressesNF-κB activation in various cancer cells (21-23). Therefore,to examine whether curcumin inhibits OPN-induced NF-κB-DNA binding, cells were pretreated with curcumin and thentreated with OPN, and EMSA was performed. The datashowed that curcumin significantly suppressed OPN-inducedNF-κB-DNA binding (Fig. 1B).

The ATF-4 response element has been reported to be inthe promoter region of VEGF (28,29). ATF-4 plays a crucialrole in OPN-induced VEGF expression (16). To confirm theeffect of curcumin on OPN-induced ATF-4-DNA binding

CHAKRABORTY et al: CURCUMIN INHIBITS OSTEOPONTIN-INDUCED BREAST TUMOR ANGIOGENESIS642

641-646 29/7/08 11:39 Page 642

and nuclear localization, cells were pretreated with curcuminand then treated with OPN. Analysis of ATF-4-DNA bindingand nuclear localization was performed by EMSA andimmunofluorescence, respectively. The data indicate thatcurcumin inhibits OPN-induced ATF-4-DNA binding as wellas nuclear localization (Fig. 1C and D). This suggests that

curcumin curbs OPN-induced NF-κB and ATF-4 activation,and thereby alters OPN-induced VEGF expression in humanbreast cancer cells.

Curcumin inhibits OPN-induced VEGF-dependent breasttumor cell motility. To ascertain whether curcumin has anyeffect on OPN-induced breast tumor cell motility, a woundmigration assay was performed using MDA-MB-231 cells.Cells were treated with OPN, or pretreated with curcuminalone or in combination with anti-VEGF or anti-NRP-1 anti-body and then treated with OPN. Wound photographs werecaptured and analyzed, and the data represented by a bar graph(Fig. 2A and B). The data indicate that curcumin significantlyinhibits OPN-induced tumor cell motility. A low concentrationof curcumin in combination with anti-VEGF or anti-NRP-1antibody also inhibited OPN-induced cell motility; however,the percent of inhibition was higher in this condition ascompared to curcumin alone (Fig. 2A and B). These datasuggest that curcumin suppresses OPN-induced VEGF/NRP-1-dependent breast tumor cell motility.


Figure 1. (A) Curcumin inhibits OPN-induced VEGF expression. MDA-MB-231 cells were treated with OPN for 6 h, or pretreated with the indicatedconcentrations of curcumin for 2 h and then treated with OPN. The level of VEGF in the cell lysate was determined by Western blot analysis. Actin was used asthe control. (B-D) Curcumin suppresses OPN-induced NF-κB and ATF-4-DNA binding and ATF-4 nuclear localization. MDA-MB-231 cells were treated withOPN alone, or pretreated with curcumin followed by treatment with OPN. NF-κB and ATF-4-DNA binding activities were measured usng EMSA. Nuclearlocalization of ATF-4 was determined by immunofluorescence. ATF-4 was stained with Cy 2 (green), and nuclei were stained with PI (red).

Figure 2. (A and B) Suppression of OPN-induced tumor cell motility bycurcumin. MDA-MB-231 cells were pretreated with curcumin alone or incombination with anti-VEGF or anti-NRP-1 antibody, followed by treatmentwith OPN. A wound assay was performed and wound photographs werecaptured (A). Data are presented in the form of a bar graph (B). *P=0.034.

641-646 29/7/08 11:39 Page 643

Curcumin suppresses OPN-induced VEGF-dependent tumor-endothelial interaction. To examine the effect of curcumin onthe suppression of OPN-induced tumor-endothelial cell inter-action, a co-migration assay was performed using MDA-MB-231 cells and HUVEC in a modified Boyden chamber. Aftertermination of the experiments, HUVEC which had migratedfrom the upper to the lower chamber were stained withGiemsa, photographed, and counted in three high-power fields.The resulting data were represented by a bar graph (Fig. 3Aand B). The data indicate that OPN-induced endothelial cellmigration towards tumor cells was suppressed by curcumin(Fig. 3A and B). More importantly, curcumin with anti-VEGFor anti-NRP-1 antibody suppressed OPN-induced migrationof endothelial cells towards tumor cells, further suggestingthat OPN-induced tumor-derived VEGF promotes tumor-endothelial interaction and that curcumin suppresses thiseffect.

Inhibition of OPN-induced in vivo Matrigel plug angiogenesisby curcumin. To determine the role of curcumin in the sup-pression of OPN-induced in vivo angiogenesis, a Matrigel plugangiogenesis assay was performed in nude mice. Conditionedmedium (CM) collected from untreated or OPN-treatedMDA-MB-231 cells was mixed with Matrigel and injectedsubcutaneously into the ventral groin region of female nudemice as described previously (16). In separate experiments,curcumin was mixed with the CM of OPN-treated MDA-MB-231 cells along with Matrigel and then injected into mice.After 21 days, the mice were sacrificed and the Matrigel plugsphotographed (Fig. 4, panel I). The data indicate that curcuminradically suppressed OPN-induced in vivo angiogenesis ascompared to the controls (Fig. 4). The excised Matrigel plugswere analyzed histopathologically (panels II and III; magni-fication, x10 and x60) and by immunofluorescence using anti-vWF (Von Willebrand Factor, an endothelial cell marker)

and anti-pKDR antibodies. The data indicate that curcuminsignificantly suppressed OPN-induced in vivo angiogenesis(panel IV) and KDR phosphorylation. Taken together, thedata demonstrate that curcumin curtails OPN-induced in vivoangiogenesis by inhibiting KDR phosphorylation.

Discussion

In the current study, we have demonstrated that curcumininhibits OPN-induced VEGF expression, which in turnsuppresses tumor and endothelial cell motility and in vivoangiogenesis. Previously, we reported that NF-κB and ATF-4play a crucial role in OPN-induced VEGF expression (16). Inthis study, we found that curcumin suppresses OPN-inducedNF-κB and ATF-4-DNA binding and ATF-4 nuclearlocalization. Thus, curcumin abrogates OPN-induced VEGFexpression by suppressing the binding activities of NF-κBand ATF-4. Our data also indicate that OPN-induced VEGF-dependent in vivo angiogenesis was downregulated bycurcumin. This clearly suggests that curcumin acts as a potentanti-angiogenic agent in regulating OPN-induced tumorangiogenesis.

The role of OPN in tumor angiogenesis is being intenselyinvestigated (30,31). Previously, OPN was considered to be ametastasis-associated gene (32). Recently, however, its rolein tumor angiogenesis has been demonstrated (16,27,30).Angiogenesis-based therapy has emerged as one of the majortherapeutic approaches to cancer (33). Thus, identifying noveltherapeutic approaches to targeting the angiogenic pathwayscould prove to be promising for cancer therapy. We previouslyshowed that OPN acts as a potent angiogenic molecule byregulating VEGF expression in a breast cancer model (16).Consequently, targeting OPN-regulated angiogenic pathwayscould be a promising therapeutic strategy for the treatment ofbreast cancer.


Figure 3. (A and B) Curcumin curbs OPN-induced endothelial cell migrationtowards the tumor cells. MDA-MB-231 cells were used in the lower chamberwhile endothelial cells were seeded in the upper chamber. MDA-MB-231cells were pretreated with curcumin alone or in combination with anti-VEGFor anti-NRP-1 antibody, then treated with OPN. A co-migration assay wasconducted. Endothelial cells which had migrated from the upper to the lowerchamber were stained with Giemsa and photographed (A), then counted inthree high-power fields. Data are presented in the form of a bar graph (B).*P=0.008, **P<0.035.

641-646 29/7/08 11:39 Page 644


Figure 4. Curcumin abrogates OPN-induced in vivo tumor angiogenesis. Typical photographs of in vivo Matrigel plug angiogenesis in nude mice. Angiogenicblood vessels are indicated by arrows. Note the significant reduction in blood vessel formation in the plug treated with curcumin and OPN compared withOPN alone (panel I). Matrigel plugs were analyzed histopathologically by H&E staining, and photographs were captured at a magnification of x10 and x60(panels II and III). Matrigel plug sections were immunohistochemically analyzed using anti-vWF and anti-pKDR antibodies (panel IV). vWF was stained withCy3 (red) and p-KDR with FITC (green). Nuclei were stained with DAPI (blue).

Figure 5. Schematic representation of the curcumin suppression of OPN-induced VEGF expression leading to the inhibition of tumor growth and angiogenesis.

641-646 29/7/08 11:39 Page 645

Curcumin has long been used as a potent anti-inflammatoryagent. Recently, though, this molecule has been identified ashaving anti-cancer properties (17-19). Curcumin also inhibitsPMA-induced VEGF mRNA expression and suppresses angio-genesis (19). Previously, we found that it suppressed OPN-regulated murine melanoma growth by inhibiting MMP-2 acti-vation (25). However, the role of curcumin in OPN-inducedbreast tumor angiogenesis is not well defined. In this study,we demonstrated that curcumin abrogates OPN-induced VEGFexpression, thus inhibiting tumor angiogenesis. Moreover,anti-VEGF or anti-NRP-1 antibody along with curcuminresulted in higher anti-angiogenic activity than curcuminalone. We therefore hypothesize that curcumin along with anti-VEGF or anti-NRP-1 antibody could be a novel approach forthe treatment of cancer.

In this study, we have gone some way in demonstratingfor the first time that curcumin inhibits OPN-induced NF-κB/ATF-4-dependent VEGF expression, ultimately suppressingbreast tumor angiogenesis (Fig. 5). As OPN is considered to bea major molecule in the control of breast cancer progression,targeting the OPN-regulated signaling pathway by curcumincould be an emerging approach to the treatment of the disease.This approach may help block the OPN-regulated VEGF-dependent angiogenic switch, and could also contribute to thedevelopment of a novel strategy for the treatment of breastcancer.

Acknowledgements

This work was supported by the Council of Scientific andIndustrial Research, Government of India (to G.C. and S.J.).

References

1. Chatterjee SK and Zetter BR: Cancer biomarkers: knowing thepresent and predicting the future. Future Oncol 1: 37-50, 2005.

2. Payne SJ, Bowen RL, Jones JL and Wells CA: Predictive markersin breast cancer - the present. Histopathology 52: 82-90, 2008.

3. Rodrigues LR, Teixeira JA, Schmitt FL, Paulsson M andLindmark-Mänsson H: The role of osteopontin in tumorprogression and metastasis in breast cancer. Cancer EpidemiolBiomarkers Prev 16: 1087-1097, 2007.

4. Mirza M, Shaughnessy E, Hurley JK, Vanpatten KA, Pestano GA,He B and Weber GF: Osteopontin-c is a selective marker ofbreast cancer. Int J Cancer 122: 889-897, 2008.

5. Rudland PS, Platt-Higgins A, El-Tanani M, De Silva Rudland S,Barraclough R, Winstanley JH, Howitt R and West CR: Prognosticsignificance of the metastasis-associated protein osteopontin inhuman breast cancer. Cancer Res 62: 3417-3427, 2002.

6. Tuck AB, Chambers AF and Allan AL: Osteopontin over-expression in breast cancer: knowledge gained and possibleimplications for clinical management. J Cell Biochem 102:859-868, 2007.

7. Jain S, Chakraborty G, Bulbule A, Kaur R and Kundu GC:Osteopontin: an emerging therapeutic target for anticancertherapy. Expert Opin Ther Targets 11: 81-90, 2007.

8. Folkman J: Fundamental concepts of the angiogenic process.Curr Mol Med 3: 643-651, 2003.

9. Carmeliet P and Jain RK: Angiogenesis in cancer and otherdiseases. Nature 407: 249-257, 2000.

10. Jain RK: Molecular regulation of vessel maturation. Nat Med 9:685-693, 2003.

11. McMahon G: VEGF receptor signaling in tumor angiogenesis.Oncologist 5: 3-10, 2000.

12. Glade-Bender J, Kandel JJ and Yamashiro DJ: VEGF blockingtherapy in the treatment of cancer. Expert Opin Biol Ther 3:263-276, 2003.

13. Jain RK: Tumor angiogenesis and accessibility: role of vascularendothelial growth factor. Semin Oncol 29: 3-9, 2002.

14. Carmeliet P: VEGF as a key mediator of angiogenesis in cancer.Oncology 69: 4-10, 2005.

15. Cross MJ, Dixelius J, Matsumoto T and Claesson-Welsh L:VEGF-receptor signal transduction. Trends Biochem Sci 28:488-494, 2003.

16. Chakraborty G, Jain S and Kundu GC: Osteopontin promotesvascular endothelial growth factor-dependent breast tumor growthand angiogenesis via autocrine and paracrine mechanisms.Cancer Res 68: 152-161, 2008.

17. Shishodia S, Chaturvedi MM and Aggarwal BB: Role of curcuminin cancer therapy. Curr Probl Cancer 31: 243-305, 2007.

18. Aggarwal BB, Sundaram C, Malani N and Ichikawa H:Curcumin: the Indian solid gold. Adv Exp Med Biol 595: 1-75,2007.

19. Arbiser JL, Klauber N, Rohan R, van Leeuwen R, Huang MT,Fisher C, Flynn E and Byers HR: Curcumin is an in vivoinhibitor of angiogenesis. Mol Med 4: 376-383, 1998.

20. Shao ZM, Shen ZZ, Liu CH, Sartippour MR, Go VL, Heber Dand Nguyen M: Curcumin exerts multiple suppressive effects onhuman breast carcinoma cells. Int J Cancer 98: 234-240, 2002.

21. Dorai T and Aggarwal BB: Role of chemopreventive agents incancer therapy. Cancer Lett 215: 129-140, 2004.

22. Aggarwal BB, Takada Y and Oommen OV: From chemo-prevention to chemotherapy: common targets and commongoals. Expert Opin Investig Drugs 13: 1327-1338, 2004.

23. Bachmeier BE, Mohrenz IV, Mirisola V, Schleicher E, Romeo F,Höhneke C, Jochum M, Nerlich AG and Pfeffer U: Curcumindown-regulates the inflammatory cytokines CXCL1 and -2 inbreast cancer cells via NFkappaB. Carcinogenesis 29: 779-789,2008.

24. Philip S and Kundu GC: Osteopontin induces nuclear factorkappa B-mediated promatrix metalloproteinase-2 activationthrough I kappa B alpha/IKK signaling pathways, and curcumin(diferulolylmethane) down-regulates these pathways. J BiolChem 278: 14487-14497, 2003.

25. Philip S, Bulbule A and Kundu GC: Osteopontin stimulatestumor growth and activation of promatrix metalloproteinase-2through nuclear factor-kappa B-mediated induction of membranetype 1 matrix metalloproteinase in murine melanoma cells. JBiol Chem 276: 44926-44935, 2001.

26. Chakraborty G, Rangaswami H, Jain S and Kundu GC: Hypoxiaregulates cross-talk between Syk and Lck leading to breast cancerprogression and angiogenesis. J Biol Chem 281: 11322-11331,2006.

27. Jain S, Chakraborty G and Kundu GC: The crucial role of cyclo-oxygenase-2 in osteopontin-induced protein kinase C alpha/c-Src/IkappaB kinase alpha/beta-dependent prostate tumorprogression and angiogenesis. Cancer Res. 66: 6638-6648, 2006.

28. Roybal CN, Yang S, Sun CW, Hurtado D, Vander Jagt DL,Townes TM and Abcouwer SF: Homocysteine increases theexpression of vascular endothelial growth factor by a mechanisminvolving endoplasmic reticulum stress and transcription factorATF4. J Biol Chem 279: 14844-14852, 2004.

29. Roybal CN, Hunsaker LA, Barbash O, Vander Jagt DL andAbcouwer SF: The oxidative stressor arsenite activates vascularendothelial growth factor mRNA transcription by an ATF4-dependent mechanism. J Biol Chem 280: 20331-20339, 2005.

30. Chakraborty G, Jain S, Behera R, Ahmed M, Sharma P, Kumar Vand Kundu GC: The multifaceted roles of osteopontin in cellsignaling, tumor progression and angiogenesis. Curr Mol Med 6:819-830, 2006.

31. Bellahcène A, Castronovo V, Ogbureke KU, Fisher LW andFedarko NS: Small integrin-binding ligand N-linked glyco-proteins (SIBLINGs): multifunctional proteins in cancer. NatRev Cancer 8: 212-226, 2008.

32. Weber GF: The metastasis gene osteopontin: a candidate targetfor cancer therapy. Biochim Biophys Acta 1552: 61-85, 2001.

33. Ruegg C and Mutter N: Anti-angiogenic therapies in cancer:achievements and open questions. Bull Cancer 94: 753-762, 2007.


641-646 29/7/08 11:39 Page 646

Documents

Science is the great antidote to the poison of enthusiasm ...shodhganga.inflibnet.ac.in/bitstream/10603/2604/14/14_appendix.pdfConference, Lonavala (MH) January 8-11th 2007. (Awarded