www.elsevier.com/locate/yexcr

Experimental Cell Research 299 (2004) 91–100

DU145 human prostate carcinoma invasiveness is modulated by

urokinase receptor (uPAR) downstream of epidermal growth

factor receptor (EGFR) signaling

Asmaa Mamoune,a,b,1 Jareer Kassis,a,b,1,2 Sourabh Kharait,a,b Susanne Kloeker,c

Elisabeth Manos,c David A. Jones,c and Alan Wellsa,b,*

aDepartment of Pathology, University of Pittsburgh, Pittsburgh, PA 15261, USAbPittsburgh VAMC, Pittsburgh, PA 15261, USA

cDivision of Molecular Pharmacology, Huntsman Cancer Institute, University of Utah, Salt Lake City, UT 84112, USA

Received 16 January 2004, revised version recieved 6 May 2004

Available online 17 June 2004

Abstract

Tumor cell motility and invasion have been linked to upregulated signaling from both the epidermal growth factor receptor (EGFR) and

that for urokinase-type plasminogen activator (uPAR). However, we do not know whether these events are interdependent or unrelated,

despite the obvious diagnostic and therapeutic implications. Gene microarray analyses have suggested that EGFR signaling via

phospholipase C-g (PLCg) induces uPAR transcription. We utilized two sublines of the DU145 human prostate carcinoma cell line that are

genetically engineered to differentially activate the EGFR/PLCg cascade and are variously invasive in vitro and in vivo. uPAR protein levels

in these cells were found to be dependent on PLC signaling, pharmacologic inhibition of PLC signaling reduced uPAR expression. To

determine whether uPAR was a required element in EGFR-mediated invasion, we stably expressed uPAR cDNA in either sense or antisense

orientation in the two DU145 sublines. Interestingly, uPA production was modulated in parallel, although to a lesser degree, with uPAR in

these sublines. Antisense to uPAR significantly restricted invasion of the highly invasive DU145 WT cells through Matrigel and reduced

aggressiveness of tumors in nude mice. Up-regulation of uPAR significantly increased the invasiveness of the moderately invasive DU145

parental (DU145 P) cells through Matrigel, but this increased invasiveness was not seen in mice. uPA activity appears to contribute to

invasiveness at least through Matrigel, as antibody to uPA or amiloride limited the transmigration. These results support a model of tumor

invasion promoted by autocrine EGFR signaling involving reinforcing altered gene expression, of uPAR at least, that further induces cell

motility. Herein, a number of key molecules whose expression levels are interrelated, including both EGFR and uPAR, are required but none

are sufficient in the absence of other keys molecules in promoting tumor progression.

D 2004 Elsevier Inc. All rights reserved.

Keywords: Tumor progression; EGF receptor; Phospholipase C-g; Motility; Migration

Introduction have elucidated a variety of cellular pathways and mecha-

One of the rate-limiting steps of prostate tumor cell

invasion is its ability to breach an extracellular matrix

(ECM) [1]. An increasing amount of recent data not only

0014-4827/$ - see front matter D 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.yexcr.2004.05.008

* Corresponding author. Department of Pathology, University of

Pittsburgh, 713 Scaife Hall, Pittsburgh, PA 15261. Fax: +1-412-647-8567.

E-mail address: wellsa@msx.upmc.edu (A. Wells).1 These authors contributed equally to this publication.2 Current address: Building 10, Room B1B40, National Cancer

Institute, NIH, Bethesda, MD 20892.

nisms involved, but also have alluded to the interdepen-

dence and cross-communication of many of these pathways.

However, one relatively common aspect of cellular invasion

mechanisms is their mediation by surface receptors which,

in transformed cell, are often overexpressed and upregu-

lated. Of these receptors, the epidermal growth factor

receptor (EGFR) is the most frequently upregulated in

tumors including prostate carcinomas [2,3]. EGFR signaling

promotes tumor progression via both epigenetic events such

as de-adhesion and cytoskeletal reorganization [4], as well

as altering the transcriptional profile of the cancer cell [5,6].

A. Mamoune et al. / Experimental Cell Research 299 (2004) 91–10092

A second, distinct class of receptor, the urokinase-type

plasminogen activator (uPAR) is similarly often found

upregulated in invasive tumor cells [7,8]; this also includes

prostate carcinomas [9–11]. uPAR is considered to promote

tumor invasion both via matrix degradation by its protease

ligand, uPA, and cellular signaling of migration via other

surface receptors including EGFR [12–14]. While both

EGFR and uPAR appear to promote tumor progression,

the extent of their interrelationship in this process is not well

characterized.

These two receptor systems are linked at many levels.

Part of cellular signaling from uPAR appears to occur via

EGFR transactivation [13,15]. Inhibition of EGFR kinase

activity blocks uPAR-initiated activation of ERK MAP

kinase but not of the small rho GTPases. As ERK is required

for EGFR induction of both motility and proliferation, it is

not surprising that abrogation of this secondary effector

blocks cell motility and proliferation induced by uPAR

[13,15]. In tumor model systems, abrogation of EGFR

signaling blocks uPAR-associated invasiveness through an

extracellular matrix [16] and growth of tumors in animals

[13,17]. Thus, EGFR appears to be a necessary element for

uPAR-mediated tumor progression.

Interestingly, transcription and production of both uPAR

and its ligand, uPA, are increased by EGFR signaling

[18,19]. Treatment of PC3 human prostate cells with genis-

tein, a tyrosine kinase inhibitor, reduces both uPAR and uPA

levels [20]. Furthermore, we found that transcription of

uPAR was increased in a highly invasive subline of

DU145 which presents upregulation of the motility-associ-

ated EGFR/phospholipase C-g (PLCg) signaling cascade

[6]. Thus, while EGFR activity increases the potential uPAR

signaling capacity, it is still not known if EGFR-mediated

tumor progression depends on uPAR upregulation and

activation.

To this end, we constructed a full-length uPAR cDNA

fragment and inserted it in both sense and antisense posi-

tions into a vector carrying a constitutively active promoter

and introduced these constructs into our DU145 sublines

that show moderate and high invasiveness [21]. We dem-

onstrated that elevating uPAR expression in parental

DU145 (DU145 P) cells increased their invasiveness

through Matrigel to levels similar to the EGFR overexpress-

ing DU145 WT cells, while the highly invasive DU145 WT

cells’ invasiveness was retarded by the antisense uPAR

cDNA fragment. To establish whether this behavior is

reproducible in vivo, we injected the highly invasive

DU145 WT cells expressing antisense uPAR cDNA into

athymic nude mice, and while control DU145 WT cells

formed highly invasive tumors, the antisense-carrying cells

were significantly less aggressive and were reduced in

diaphragm invasiveness. However, upregulation of uPAR

in parental DU145 cells did not significantly increase

invasion in vivo. These results indicate that uPAR is

required, but not sufficient for EGFR-induced tumor cell

invasion.

Materials and methods

Animals

Male athymic BALB/c nu/nu mice, 6–8 weeks of age,

were purchased from the Animal Production Area of the

National Cancer Institute-Frederick Cancer Research and

Development (Frederick, MD) and housed in suitable con-

ditions. Mouse weights ranged from 20–30 g at onset of

experiments. All animal experimentation was approved by

the Institutional Animal Care and Use Committees of both

the University of Pittsburgh and the Pittsburgh VA Medical

Center.

Cell culture

DU145 human prostate carcinoma cells were derived as

described [22]; we designate these cells as parental

(DU145 P). DU145 WT prostate carcinoma cells were

generated as described previously by retroviral transduction

[23]. Two independent polyclonal isolates were used

throughout the experiments. WT cells overexpress EGFR

to levels that oversaturate the capacity of the degradation

pathway and therefore do not undergo autocrine-induced

downregulation. The DU145 WT cells present approxi-

mately one-third more EGFR when autocrine downregula-

tion is minimized and over twice the number of bindings in

the face of autocrine downregulation [23]. The cells were

maintained in DMEM (4.5 g/ml glucose) containing 10%

FBS and supplemented with L-glutamine (2 mM), penicil-

lin/streptomycin (100 units/ml), nonessential amino acids

(0.1 mM), and sodium pyruvate (1 mM). For stable

selection of WT EGFR, cells were grown in G418 (1000

Ag/ml). Before in vitro experimentation, cells were qui-

esced in DMEM containing 0.1% dialyzed FBS for 24 h.

Sense and antisense uPAR constructs were introduced

into the cells by electroporation. Briefly, uPAR was cloned

into the constitutively active pXf vector, which contains the

SV40 early promoter [24], in either sense or antisense

directions. The sense uPAR gene was cloned via PCR with

the XbaI restriction site at the 5V end and the HindIII site at

the 3V end, while the antisense gene was cloned by switch-

ing the positions of these two restriction sites. Primers used

for the sense constructs were as follows: leading strand

5VGGA TCT AGA CCA TGG GTC ACC CGC CGC,

lagging strand 5V GGA AGC TTC TAT CAG CGG TAC

TGG ACATCC AG. To append the GPI anchor to the sense

construct, two-step PCR was performed using the leading

strand primer above and 5V GGT GAT GGT GAG GCT

GAG ATG GGC AGG GCC AGG CTG AGG AGC AGC

CCC ACT GCG GTA CTG GAC ATC C for the lagging

strand, followed by PCR of the resulting template with the

same leading strand primer and 5V GGC AAG CTT TTA

GGT CCA GAG GAG AGT GCC TCC CCA CAG TCT

GGC AGT CAT TAG CAG GGT GAT GGT GAG GCT G

as the lagging strand primer. Primers for the antisense

A. Mamoune et al. / Experimental Cell Research 299 (2004) 91–100 93

construct were as follows: leading strand 5V GGA TCT

AGATCA GCG GTA CTG GAC ATC CAG, lagging strand

5V GGA AGC TTA CCA TGG GTC ACC CGC CGC.

Stable polyclonal transfectant cells, representing at least 20

initial colonies, were selected by additionally supplementing

the above media with 1.2 Ag/ml methotrexate. Two inde-

pendent polyclonal isolates were tested for each construct

with similar results.

Immunoblotting

Cells were grown in six-well plates to semiconfluence,

washed with PBS, and lysed with Laemmli buffer. The

lysate was separated by SDS–PAGE, transferred to a nylon

membrane, and immunoblotted. Primary antibodies used

were anti-uPAR, which binds to both uPAR and uPA/uPAR

complexes (American Diagnostica; catalog #3936), anti-

EGFR (Zymed; catalog #28-0005), anti-uPA (Oncogene,

catalog #IM13L), and antiactin (Sigma; catalog #A 2066).

The staining was visualized by a secondary antibody linked

to horseradish peroxidase and detected by ECL (Amer-

sham). Blots were scanned and quantitated using NIH

Image software.

Immunofluorescence

Cells, grown on coverslips, were fixed in 3% parafor-

maldehyde for 30 min and washed twice with PBS. In a

series of experiments, before fixation, the cells were treated

with or without bacterial PI-PLC, 0.5 U/ml for 1 h (Sigma;

catalog #8804) to remove any GPI-anchored uPAR. The

coverslips were blocked with 1% BSA for 1 h and then

stained with the anti-uPAR antibody (1:50 in PBS contain-

ing 1% BSA) for 2 h, washed twice in PBS/BSA, and

visualized with FITC-conjugated antimouse IgG (1:500 in

PBS/BSA) (Molecular Probes).

Transmigration assays

Cell invasiveness in vitro was determined by the ability

to transmigrate a layer of ECM, Matrigel, in a modified

Boyden chamber assay. Matrigel invasion chamber plates

were obtained from Becton Dickinson/Biocoat (catalog

#40480). For each experiment, 20,000 cells were plated

randomly and distributed among plates with different lot

numbers, with each experiment performed in duplicate or

triplicate. In one series of experiments, we incorporated into

the assay either antibody to uPA (4 Ag/ml; Oncogene,

catalog #IM13L) or amiloride (0.5 mM; Sigma) to block

the functioning of uPA. Despite possible variances in EGFR

ligand concentrations in Matrigel, these intrinsic concen-

trations are saturating since even ‘‘reduced growth factor

Matrigel’’ contains relatively high amounts of EGF (up to

0.5 ng/ml compared to 0.5–1.3 ng/ml in regular Matrigel)

(Becton Dickinson Labware 1997/98 catalog description,

page 128). Cells were kept in serum-free media containing

1% BSA for the first 24 h and then replaced with serum-free

media for the remaining 24 h. The bottom, targeting well

was filled with media containing 10% FBS. Enumeration of

the cells that invaded through the matrix over a 48-h period

was accomplished by removing uninvaded cells with a

cotton swab and then fixing, staining with crystal violet,

and visually counting cells on the bottom of the membrane.

Tumor cell inoculations

We utilized the intraperitoneal xenograft model for

invasiveness as determined by diaphragmatic invasion

[21,25]. This mode of invasion through the diaphragm

was chosen for reproducibility and quantitation; it correlates

strongly with invasion from orthotopic tumor growth. Cells

were suspended in culture media and injected (26.5-gauge

needle) into the peritoneal cavity (2 million cells in 200-Altotal volume). Mice were asphyxiated after 60 days with

halothane and necropsied with removal of the pancreas,

spleen, kidneys, diaphragm, lungs, and any tumorous tissue.

All tissues were fixed in 10% buffered formalin, paraffin-

embedded, serially sectioned, and stained with hematoxylin

and eosin (H&E). Invasion of the diaphragm was deter-

mined by microscopic quantitation. The protocol was ap-

proved by the Pittsburgh Veteran Administration Medical

Center Institutional Animal Care and Use Committee.

Results

uPAR protein levels are upregulated in invasive WT DU145

cells in relation to PLC signaling

Cell lysates of the two DU145 sublines were queried for

uPAR levels. As shown in Fig. 1A, uPAR protein levels

were about two times higher in DU145 WT overexpressing

EGFR receptor than parental DU145 cells (2.04 F0.49, N =

4, P < 0.01); the increased level of EGFR is consistent with

earlier studies, with the slight electrophoretic retardation

reflecting the reported greater stoichiometry of phosphory-

lation [23]. Surface location of the uPAR was confirmed by

treating the cells with bacterial PI-PLC to cleave the GPI

anchors of uPAR; this treatment removed much of the uPAR

staining (Fig. 1B). This level of increase was in the range

expected from an earlier gene microarray query that found

higher uPAR mRNA in DU145 WT compared to DU145 P

cells [6].

To investigate the causal link between the EGFR/PLCg

and uPAR signaling, we treated the highly invasive cell line

either with an EGFR tyrosine kinase inhibitor, PD153035, or

the specific PLC inhibitor, U73122. Exposure of DU145WT

cells to U73122 reduced significantly the level of uPAR

protein (68%F 2.5% decrease compared to nontreated cells,

n = 3), while no effect was seen using its inactive analogue,

U73343 (Fig. 1C). The EGFR kinase inhibitor PD153035

also decreased uPAR expression (37% F 12% decrease).

Fig. 1. Expression of endogenous uPAR protein in DU145 sublines. (A) Whole cell lysates from each of DU145 P and DU145 WT cells were collected and

quantitated using the BioRad protein assay. Identical concentrations of lysates were then separated in 10% SDS–PAGE and blotted with anti-uPAR antibody

under nonreducing conditions (band at 50 kDa). Concurrent blotting of an identical gel with antiactin (50 kDa) (for gel loading) or anti-EGFR (170 kDa) was

performed. Shown are representative of four independent blots in A. (B) DU145 sublines were examined for cell surface expression of uPAR by

immunofluorescence. Alternate coverslips were treated with bacterial PI-PLC to remove GPI anchors. Exposure times were preset to compare treated and

untreated sublines but were not constant between the DU145 P and DU145 WT sublines. Shown are representative of three independent analyses; uPAR is

green, red is nuclear staining. (C) Lysates from DU145 WT cells treated with 500 nM PD153035, 2 AM U73122, or 2 AM U73343 for 24 h were collected and

subjected to SDS–PAGE electrophoresis. Three different blots were scanned, the ratio uPAR/actin was determined using NIH Image software, and values were

normalized to untreated cells. Shown is a representative immunoblot of three.

A. Mamoune et al. / Experimental Cell Research 299 (2004) 91–10094

EGFR inhibition did not suppress uPAR as extensively as that

due to U73122, suggesting that other signals also are increas-

ing uPAR via PLC-mediated transcription.

Modulation of uPAR expression alters invasiveness through

Matrigel

The above studies suggested that uPAR expression may

either enhance invasion or be required for invasion. To

determine the requirement for uPAR in DU145 tumor cell

invasion, we selected DU145 WT cells since these cells are

significantly more invasive than parental cells both in vitro

and in vivo [21,23]. To determine if uPAR expression

enhances invasiveness, we utilized the moderately invasive

DU145 P line.

To study the effects of uPAR downregulation in aggres-

sively invasive cells, we cloned a full-length uPAR cDNA

fragment (including the GPI anchoring domain) downstream

from a constitutively active SV40 early promoter in the

eukaryotic expression vector pXf [25]. Stable selection of

the electroporated construct was accomplished using 1.2 Ag/ml methotrexate in the media. As shown in Fig. 2A, immu-

noblot analysis of cell lysates indicated increased expression

of uPAR protein in the sense construct-expressing cells (1.87

F 0.37, N = 4, P = 0.01). Similarly, we cloned the uPAR

cDNA fragment in the antisense direction into the pXf

promoter and stably selected the construct in the highly

invasive DU145WTcells; uPAR protein levels were reduced

significantly in the DU145 WTcells expressing the antisense

cDNA (0.38 F 0.11, N = 3, P < 0.01). Both DU145 P and

DU145WT lines were transfected and stably selected with an

empty pXf vector and ones containing irrelevant cDNA;

examination of lines demonstrated no change in uPAR levels

(data not shown). uPAR was visualized on the transfected

cells demonstrating altered levels in nearly all the cells and

that the majority of the staining was indicative of cell surface

uPAR. These findings are significant to our study since uPAR

is often actively cleaved at the GPI anchor and shed [26],

thereby reducing levels at the cell surface.

The second aspect of the uPAR system is its ligand, the

protease uPA. uPA has been reported as upregulated in

response to EGFR signaling, including in prostate carcino-

ma cells [16–19]. We found that uPA levels changed

qualitatively in the same manner as uPAR (Fig. 2B),

although to a lesser quantitative extent. It is possible that

the change in uPA levels might reflect the changes in surface

uPAR binding. Removal of the surface associated-uPAR

with bacterial PLC did not eliminate the differences noted

Fig. 2. Expression of sense or antisense uPAR cDNA constructs alters uPAR levels in DU145 sublines. (A) DU145 WT cells expressing antisense, DU145 P

overexpressing uPAR cDNA, or mock-transfected cells were lysed in equal protein contents after counting cells in separate identical wells and normalizing

protein concentrations. Lysates were separated by SDS–PAGE (10% gel) under nonreducing conditions and immunoblotted with anti-uPAR antibody. (B)

DU145 WT cells expressing antisense uPAR, DU145 P overexpressing uPAR cDNA or mock-transfected cells were lysed in equal protein contents after

counting cells in separate identical wells and normalizing protein concentrations. Lysates were separated by SDS–PAGE (10% gel) and immunoblotted with

anti-uPA antibody. (C) Protein expression of uPAwas determined in DU145 P and DU145 WT cells stably expressing the sense or antisense uPAR constructs.

The cells were lysed and subjected to immunoblot detection before or after treatment with bacterial PI-PLC to remove GPI anchors and uPAR-complexed uPA.

uPAR was also visualized to demonstrate enzymatic removal of surface uPAR, denoted by decrease in total uPAR levels. Shown are representative experiments

of four (A) or two (B and C).

A. Mamoune et al. / Experimental Cell Research 299 (2004) 91–100 95

(Fig. 2C). More importantly than the change in uPA levels

which are slight compared to alterations in uPAR, these

immunoblots demonstrate that both uPAR and uPA are

produced by these cells, allowing for a human uPA/human

uPAR autocrine loop.

A hallmark of tumor progression is its ability to invade

through the ECM into surrounding tissue. We asked whether

lowering uPAR protein levels in the highly invasive DU145

WT cells reduced these cells invasiveness. To this end, we

compared transmigration of DU145 WT and DU145 WTAS

cells through Matrigel; transmigration by DU145 WT AS

was drastically reduced compared to the mock-transfected

cells (Fig. 3A). These results clearly indicated that down-

regulating uPAR reversed the invasive properties of upregu-

lated EGFR in these cells. Along the same lines, we

overexpressed uPAR in the less invasive parental DU145

cell lines (DU145 P SN). As shown in Fig. 3B, DU145 P

SN transmigration through Matrigel was dramatically in-

creased compared to DU145 P cells; uPAR expression

compensated for the relatively less PLCg signaling in these

cells [25]. Again, these results indicate that uPAR expres-

sion levels are both required for and contribute to invasion

at least in vitro and thus confirm a rate-limiting role for

uPAR functions.

uPAR initiates signals upon binding its ligand [7,8].

However, there have been suggestions of ligand-indepen-

dent uPAR modulation of cell behaviors. Addition of an

antibody against human uPA reduced DU145 WT invasive-

ness by about two thirds indicating that the invasiveness

noted herein is ligand-dependent (Fig. 3C). It is a further

question whether uPA protease activity is required for

uPAR-enhanced motility [27–29]. Amiloride binds to and

inhibits human uPA [30], and has been used to limit

invasiveness in vitro and in vivo of human carcinoma lines

[31,32]. We find that DU145 WT invasion of the Matrigel

barrier is partly reduced by amiloride (Fig. 3C), suggesting a

role for proteolytic activity.

uPAR downregulation reduces tumor formation and

invasiveness in mice

Our data so far showed a clear role for uPAR and uPA in

tumor cell invasion through Matrigel. In our previous

studies, we showed that DU145 invasiveness correlated

with PLCg activation via overexpressed EGFR. Inhibition

of PLCg pharmacologically with U73122 [21] or molecu-

larly via the dominant negative fragment PLCz [25] signif-

icantly reduced invasion of tumors formed by DU145 WT in

athymic nude mice. Because our recent data showed a clear

link between enhanced invasiveness and uPAR expression,

we asked whether downregulation of uPAR expression

would negate the effects of upregulated EGFR and PLCg

activity. Being able to demonstrate such an effect in mice

would be of paramount importance due to the fact that the

complex interactions of the uPA/uPAR system with various

integrins, growth factor receptors, and extracellular pro-

Fig. 3. Invasion of DU145 cells is influenced by uPAR expression. (A) Antisense uPAR reduced invasion of DU145 WT cells. (B) Sense uPAR increased the

invasiveness of DU145 P cells. (C) Anti uPA antibody reduced invasion of DU145 WT by two thirds and amiloride treatment decreased DU145 WT invasion

by about one third. Twenty thousand cells were plated into Matrigel-coated filters in transwell chambers. After 48 h, the transmigrated cells were quantitated by

manual counting under a microscope. Cells expressing sense or antisense uPAR cDNA (solid bars in A and B); DU145 cells not expressing uPAR constructs

(open bars in A and B). Shown are the mean F SEM of 3 (A and B) or 2 (C) experiments each in triplicate or duplicate. *P < 0.05 compared to mock-

transfected (A and B) or diluent (C).

A. Mamoune et al. / Experimental Cell Research 299 (2004) 91–10096

teases can only be fully realized in an in vivo environment

as opposed to the controlled Matrigel. This is possible

because the cells express both the receptor, uPAR, and the

ligand, uPA, potentially providing for autocrine activation.

We injected both DU145 WT and DU145 WT AS cells

intraperitoneally into athymic nude mice (approximately 2

million cells each) and allowed tumors to form over a 60-

day period. DU145 WT cells aggressively invaded through

the diaphragm of these mice as expected [21] (Fig. 4).

However, DU145 WT AS cells formed markedly less

aggressive tumors in these mice (Table 1). Upregulation of

uPAR in the moderately invasive parental DU145 cells was

tested for invasion in vivo. The overexpression of uPAR in

these cells did not increase the invasion into the diaphragm

significantly (Table 1), although greater microinfiltration of

individual tumor cells could be noted in some of the tumors

(Fig. 4). While this lack of significantly increased invasion

is seemingly at odds with the enhanced Matrigel transmi-

gration, actual tumor cell progression obviously requires

properties in addition to those needed for in vitro transmi-

gration. In sum, these data strongly support a critical role for

uPAR in realizing the invasive potential driven by EGFR/

PLCg signaling.

Discussion

The signals promoting tumor cell invasion and metas-

tasis are intertwined in a complex medley of signaling

from growth factors and ECM components [1]. In this

study, we probed downstream of a previously established

finding that growth factor-induced migration, via PLCg

signaling, is rate-limiting in tumor cell invasion [33].

While PLCg contributes to motility induced by numerous

growth factor receptors [34–36] via an epigenetic hydro-

lysis of PIP2 and mobilization of actin-binding proteins

[37,38], there is evidence that signaling through this

effector also alters the gene expression profile [6]. Thus,

we asked whether these changes in transcription contribute

to tumor invasion. Using mRNA microarray analysis, we

discovered a number of genes that were upregulated in the

human prostate carcinoma DU145 cell sublines in con-

junction with increased invasiveness both in vitro and in

vivo. The urokinase receptor uPAR was found to be

downregulated when the human prostate cancer line

PC3, which also presents upregulated autocrine EGFR

signaling, was challenged with the EGFR-preferential

kinase inhibitor genistein [20]. Thus, we focused on

clarifying the relationship between these two receptor

systems in promoting tumor cell progression. We showed

for the first time that EGFR signaling increases uPAR-

mediated phenotypic cell behaviors in tumor cells and that

this increase in uPAR activity is downstream of PLCg

signaling. Most importantly, using both in vitro and in

vivo models of tumor invasion, we demonstrated that the

uPAR/uPA complex contributes to invasion of these pros-

tate tumor cells but only in the presence of a functional

EGFR/PLCg signaling cascade. In short, the findings paint

a picture not of two separate and distinct regulators of

invasion but that these two critical pathways, uPAR and

EGFR, are inexorably linked. This reinforcing cascade of

invasive signaling from cell surface receptors provides two

Fig. 4. Diaphragm invasion of DU145 cells in athymic nude mice. Mice were injected in the peritoneal cavity with 2 million cells. (A) DU145 WT or DU145

WT antisense uPAR cells or (B) parental DU145 or parental DU145 expressing sense uPAR cells were inoculated. Mice were killed after 60 days, and organ/

tumor growth and metastasis were determined. Organs, including the diaphragm, were removed, fixed in 10% formalin and stained with H&E. The left column

shows a representative diaphragm taken from a mouse injected with DU145 WT or parental DU145, whereas the right column indicates diaphragms taken from

mice injected with DU145 WT antisense uPAR or parental DU145 sense uPAR cells.

A. Mamoune et al. / Experimental Cell Research 299 (2004) 91–100 97

new molecules that would be extracellularly accessible

targets for intervention.

Our previous studies demonstrated a rate-limiting role for

PLCg activity in tumor cell invasion. In prostate, breast and

bladder carcinomas, and glioblastoma, inhibition of PLCg

significantly retarded cell invasion in vitro and/or in vivo

[21,23,25,33]. In both breast and prostate cells, invasion

through PLCg activity was shown to be secondary to

autocrine EGFR activity, since (a) a truncated EGFR which

does not signal PLCg had reduced invasion [21] and (b)

Table 1

Invasion into the diaphragm of intraperitoneal tumors from DU145 sublines

Wild-type Wild-type

AS-uPAR

Parental Parental SN

Diaphragm

tumorsa24/24 17/17 16/16 14/15

Diaphragm

invasivenessb2.45 F 0.3 1.29* F 0.36 2.18 F 0.34 2.28 F 0.32

Data were collated from three different experiments of study with 5–8 mice

each.a Number of mice with tumors on diaphragm/mice inoculated.b On a scale of 0–4, with 0 being noninvasive with demonstrable fibrosis

capsule and 4 being transdiaphragmatic invasion.

*P < 0.05.

treating cells with the EGFR-specific kinase inhibitor

PD153035 likewise retarded cell invasion through Matrigel

[39]. These findings strongly suggest a pivotal role for

EGFR (or other growth factor receptors) promoting invasion

through the PLCg motility pathway. However, other signal-

ing pathways likely also contribute to invasion. Previous

studies have documented the role of upregulated uPAR in

invasion of various tumors, including glioblastomas, carci-

nomas of the bladder and breast, and highly metastatic

human lung cancer cells [10,40–42]. That this is linked to

tumor progression has been shown by interventional studies.

Antibodies against uPA or uPAR inhibit cancer invasion and

metastasis in vivo or in vitro in various tumors [11,43].

Downregulation of uPAR mRNA using anti-uPAR oligonu-

cleotides or antisense uPAR cDNA reversed the invasive

phenotype of human fibroblasts and inhibited the invasion

of human colon cancer cell lines, respectively [44]. These

tumors are the same types in which upregulated EGFR

signaling strongly correlates with invasion and tumor pro-

gression [45–47]. However, it was not probed whether these

two pathways were upregulated separately or as a concor-

dant event. Our results herein strongly suggest a causal link

between EGFR-PLCg signaling and uPAR upregulation.

Thus, it appears that these two surface receptors are coor-

A. Mamoune et al. / Experimental Cell Research 299 (2004) 91–10098

dinately upregulated during progression of human tumors.

However, it was quite unknown if the contributions of

EGFR and uPAR signaling to tumor invasion were linked

or separate.

That uPAR and, to a lesser extent, uPA were found to be

induced by autocrine EGFR signaling through the PLCg

signaling pathway strongly suggests that the invasiveness

driven by these two molecules are interlinked. Particularly

intriguing is the finding that uPAR’s role in tumor invasion

likely occurs through increased cell migration. uPAR is

known to localize at the leading edge of the cell during

the migration process [48] and has been shown to be

required for migration of a number of cells. It also is vital

for promoting integrin signaling (reviewed in [8]), along

with various integrin-linked molecules such as focal adhe-

sion kinase (FAK) and mitogen-activated protein kinase

(MAPK) [49]. More recently, uPAR has been shown to

promote extravillous trophoblast migration via the ERK

MAP kinase and PLC pathways [29]. Furthermore, these

promotility signals from uPAR may occur, at least in part,

via transactivation of EGFR [50]. These same intracellular

effectors (ERK, PLC, etc.) are also required for signaling

from EGFR and other growth factor receptors [49,51]. Thus,

coordinate upregulation of uPAR in the face of an active

TGFa-EGFR autocrine signaling loop, present in practically

all prostate carcinomas [3], would serve either to reinforce

these signaling pathways or broaden the biological

responses contingent upon these.

We found that uPAR function was rate-limiting for

invasion of these cells that have active EGFR autocrine

signaling and activation of PLCg. In vitro transmigration of

the ECM Matrigel was modulated by uPAR levels in the

highly invasive DU145 WT and moderately invasive pa-

rental DU145 sublines (Table 1); abrogating uPAR by

antisense in DU145 WT (DU145 WT AS) reduced tumor

aggressiveness, whereas upregulation in parental DU145

(DU145 P SN) did not lead to significantly greater invasion

in vivo. Thus, invasion in vivo requires properties in

addition to those needed for ECM transmigration in vitro,

or uPAR may interact with tumor cell- or organismal-

derived ligands and inhibitors that modulate uPAR/uPA

functioning in animals. One suggestive hint is the finding

that HGF is also upregulated at the transcriptional level in

these same DU145 WT cells but not in the DU145 P cells

[6], particularly so since uPA proteolytically processes

HGF to an active form. Still, whatever the reason, uPAR

overexpression in itself was not capable of restoring the

increased invasiveness noted in the DU145 WT cells in

which endogenous uPAR expression is increased. This

demonstrates that the contribution of uPAR to tumor

invasion is not an independent event but linked to a larger

web of signals. That uPAR is required even for EGFR-

mediated tumor invasion presents obvious implications for

rational therapeutic interventions in that uPAR, being on

the cell surface, is targetable by noncell permeant agents.

Still, for this approach to be realized, the operative mech-

anism of uPAR signaling (i.e., protease, complexing with

integrins, synergizing with EGFR [50], or direct receptor

signaling) in this context needs to be determined.

One last point warrants further emphasis. While uPAR

gene expression levels correlated most strongly with inva-

siveness in our DU145 cells lines, a number of other

prominent receptors or ligands also were overexpressed in

DU145 WT, including HGF, IGF-1, and the integrin subunit

av. Not only have these been implicated in promoting tumor

cell invasion, but they are also known to interact with uPAR

in promoting cell growth as well as migration and invasion

[52,53]. This is particularly relevant for HGF, which is

activated from its proform by uPA-mediated cleavage

[54]. Thus, this initial study, in only prostate cancer that

will require experimental validation in other tumor types,

may only be the tip of the iceberg with a panoply of genes

being expressed in concert that further reinforce the motility,

scatter, and proliferative properties needed for successful

access to and growth in ectopic sites that are the essence of

metastasis.

Acknowledgments

We thank Wendy Mars for helpful discussions and

guidance and Diana Whaley and James Solava for critical

technical support. This study was supported by grants from

the Veterans Adminstration Merit Award Program and the

National Cancer Institute at NIH.

References

[1] A. Wells, Tumor invasion: role of growth factor-induced cell motility,

Adv. Cancer Res. 78 (2000) 31–101.

[2] S.A. Aaronson, Growth factors and cancer, Science 254 (1991)

1146–1153.

[3] H. Kim, T. Turner, J. Kassis, J. Souto, A.Wells, EGF receptor signaling

in prostate development, Histol. Histopathol. 14 (1999) 1175–1182.

[4] A. Glading, D.A. Lauffenburger, A. Wells, Cutting to the chase:

calpain proteases in cell migration, Trends Cell Biol. 12 (2002)

46–54.

[5] A. Malliri, M. Symons, R.F. Hennigan, A.F.L. Hurlstone, R.F. Lamb,

T. Wheeler, B.W. Ozanne, The transcriptional factor AP-1 is required

for EGF-induced activation of rho-like GTPases, cytoskeletal rear-

rangements, motility, and in vitro invasion of A431 cells, J. Cell Biol.

143 (1998) 1087–1099.

[6] E.J. Manos, M. Kim, J. Kassis, B. Chang, A. Wells, D.A. Jones,

Prostin-1, a novel phospholipase C-g regulated gene negatively asso-

ciated with prostate tumor invasion, Oncogene 20 (2001) 2781–2790.

[7] P.A. Andreasen, R. Egelund, H.H. Petersen, The plasminogen activa-

tion system in tumor growth, invasion, and metastasis, Cell. Mol. Life

Sci. 57 (2000) 25–40.

[8] F. Blasi, P. Carmeliet, uPAR: a versatile signalling orchestrator, Nat.

Rev., Mol. Cell Biol. 3 (2002) 932–943.

[9] D.F. Liu, S.A. Rabbani, Induction of urinary plasminogen activator by

retinoic acid results in increased invasiveness of human prostate can-

cer cells PC-3, Prostate 27 (1995) 269–276.

[10] H. Miyake, I. Hara, K. Yamanaka, S. Arakawa, S. Kamidono, Eleva-

tion of urokinase-type plasminogen activator and its receptor densities

A. Mamoune et al. / Experimental Cell Research 299 (2004) 91–100 99

as new predictors of disease progression and prognosis in men with

prostate cancer, Int. J. Oncol. 14 (1999) 535–541.

[11] S.A. Rabbani, J. Gladu, Urokinase receptor antibody can reduce tu-

mor volume and detect the presence of occult tumor metastases in

vivo, Cancer Res. 62 (2002) 2390–2397.

[12] N. Sidenius, F. Blasi, The urokinase plasminogen activator system in

cancer: recent advances and implications for prognosis and therapy,

Cancer Metastasis Rev. 22 (2003) 205–222.

[13] D. Liu, J.A. Aguirre-Ghiso, Y. Estrada, L. Ossowski, EGFR is a

transducer of the urokinase receptor initiated signal that is required

for in vivo growth of a human carcinoma, Cancer Cell 1 (2002)

445–457.

[14] J.A. Aguirre-Ghiso, D. Liu, A. Mignatti, K. Kovalski, L. Ossowski,

Urokinase receptor and fibronectin regulate the ERKMAPK to

p38MAPK activity ratios that determine carcinoma cell proliferation

or dormancy in vivo, Mol. Biol. Cell 12 (2001) 863–879.

[15] M. Jo, K.S. Thomas, D.M. O’Donnell, S.L. Gonias, Epidermal

growth factor receptor-dependent and -independent cell-signaling

pathways originating from the urokinase receptor, J. Biol. Chem.

278 (2003) 1642–1646.

[16] A. Unlu, R.E. Leake, The effect of EGFR-related tyrosine kinase

activity inhibition on the growth and invasion mechanisms of prostate

carcinoma cell lines, Int. J. Biol. Markers 18 (2003) 139–146.

[17] T. Shiratsuchi, H. Ishibashi, K. Shirasuna, Inhibition of epidermal

growth factor-induced invasion by dexamethasone and AP-1 decoy

in human squamous cell carcinoma cell lines, J. Cell. Physiol. 193

(2002) 340–348.

[18] P.J. Jensen, U. Rodeck, Autocrine/paracrine regulation of keratinocyte

urokinase plasminogen activator through the TGF-alpha/EGF recep-

tor, J. Cell. Physiol. 155 (1993) 333–339.

[19] D.F. Jarrard, B.F. Blitz, R.C. Smith, B.L. Patai, D.B. Rukstalis,

Effects of epidermal growth factor on prostate cancer cell line PC3

growth and invasion, Prostate 24 (1994) 46–53.

[20] Y. Li, F. Sarkar, Down-regulation of invasion and angiogenesis-related

genes identified by cDNA microarray analysis of PC3 prostate cancer

cells treated with genistein, Cancer Lett. 186 (2002) 157–164.

[21] T. Turner, P. Chen, L.J. Goodly, A. Wells, EGF receptor signaling

enhances in vivo invasiveness of DU-145 human prostate carcinoma

cells, Clin. Exp. Metastasis 14 (1996) 409–418.

[22] K. Stone, D.D. Mickey, H. Wunderli, G.H. Mickey, D.F. Paulson,

Isolation of a human prostate carcinoma cell line (DU145), Int. J.

Cancer 21 (1978) 274–281.

[23] H. Xie, T. Turner, M.-H. Wang, R.K. Singh, G.P. Siegal, A. Wells, In

vitro invasiveness of DU-145 human prostate carcinoma cells is

modulated by EGF receptor-mediated signals, Clin. Exp. Metastasis

13 (1995) 407–419.

[24] P. Chen, K. Gupta, A. Wells, Cell movement elicited by epidermal

growth factor receptor requires kinase and autophosphorylation but is

separable from mitogenesis, J. Cell Biol. 124 (1994) 547–555.

[25] T. Turner, M. VanEpps-Fung, J. Kassis, A. Wells, Molecular inhibi-

tion of PLCg signaling abrogates DU-145 prostate tumor cell inva-

sion, Clin. Cancer Res. 3 (1997) 2275–2282.

[26] G. Hoyer-Hansen, E. Ronne, H. Solberg, N. Behrendt, M. Ploug,

L.R. Lund, V. Ellis, K. Dano, Urokinase plasminogen activator

cleaves its cell surface receptor releasing the ligand-binding domain,

J. Biol. Chem. 267 (1992) 18224–18229.

[27] D.H. Nguyen, I.M. Hussaini, S.L. Gonias, Binding of urokinase-type

plasminogen activator to its receptor in MCF-7 cells activates extra-

cellular signal-regulated kinase 1 and 2 which is required for in-

creased cellular motility, J. Biol. Chem. 273 (1998) 18268–18272.

[28] N. Busso, S.K. Masur, D. Lazega, S. Waxman, L. Ossowski, Induc-

tion of cell migration by pro-urokinase binding to its receptor:

possible mechanisms for signal transduction in human epithelial

cells, J. Cell Biol. 126 (1994) 259–270.

[29] J. Liu, C. Chakraborty, C.H. Graham, Y.P. Barbin, S.J. Dixon, P.K.

Lala, Noncatalytic domain of uPA stimulates human extravillous tro-

phoblast migration by using phospholipase C, phosphatidylinositol 3-

kinase and mitogen-activated protein kinase, Exp. Cell Res. 286

(2003) 138–151.

[30] J. Jankun, E. Skrzypczak-Jankun, Binding sites of amiloride to uro-

kinase plasminogen activator depends on species, Int. J. Mol. Med.

8 (2001) 365–371.

[31] N. Oka, Y. Okumura, H.O. Kanayama, H. Izaki, M. Okamoto, H.

Kido, S. Kagawa, Amiloride and urinary trypsin inhibitor inhibit

urothelial cancer invasion, Eur. Urol. 44 (2003) 737–741.

[32] J. Jankun, R.W. Keck, E. Skrzypczak-Jankun, R. Swiercz, Inhibitors

of urokinase reduce size of prostate cancer xenografts in severe com-

bined immunodeficient mice, Cancer Res. 57 (1997) 559–563.

[33] J. Kassis, D.A. Lauffenburger, T. Turner, A. Wells, Tumor invasion as

dysregulated cell motility, Semin. Cancer Biol. 11 (2001) 105–118.

[34] K.E. Bornfeldt, E.W. Raines, T. Nakano, L.M. Graves, E.G. Krebs, R.

Ross, Insulin-like growth factor-1 and platelet-derived growth factor-

BB induce directed migration of human arterial smooth muscle cells

via signalling pathways that are distinct from those of proliferation,

J. Clin. Invest. 93 (1994) 1266–1274.

[35] V. Kundra, J.A. Escobedo, A. Kazlauskas, H.K. Kim, S.G. Rhee, L.T.

Williams, B.R. Zetter, Regulation of chemotaxis by the platelet-de-

rived growth factor receptor-h, Nature 367 (1994) 474–476.

[36] P. Chen, H. Xie, M.C. Sekar, K.B. Gupta, A. Wells, Epidermal growth

factor receptor-mediated cell motility: phospholipase C activity is

required, but MAP kinase activity is not sufficient for induced cell

movement, J. Cell Biol. 127 (1994) 847–857.

[37] P. Chen, J. Murphy-Ullrich, A. Wells, A role for gelsolin in actua-

ting EGF receptor-mediated cell motility, J. Cell Biol. 134 (1996)

689–698.

[38] J. Chou, D. Beer-Stolz, N. Burke, S.C. Watkins, A. Wells, Distribu-

tion of gelsolin and phosphoinositol 4,5-bisphosphate in lamellipodia

during EGF-induced motility, Int. J. Biochem. Cell Biol. 34 (2002)

776–790.

[39] J. Kassis, J. Moellinger, H. Lo, N. Greenberg, H.-G. Kim, A. Wells, A

role for phospholipase C-g-mediated signaling in tumor cell invasion,

Clin. Cancer Res. 5 (1999) 2251–2260.

[40] S.J. Gong, S.Y. Rha, H.C. Chung, N.C. Yoo, J.K. Roh, W.I. Yang,

K.S. Lee, J.S. Min, B.S. Kim, H.C. Chung, Tissue urokinase-type

plasminogen activator receptor levels in breast cancer, Int. J. Mol.

Med. 6 (2000) 301–305.

[41] M. Hudson, M. McReynolds, Urokinase and the urokinase receptor:

association with in vitro invasiveness of human bladder cancer cell

lines, J. Natl. Cancer Inst. 89 (1997) 709–717.

[42] T.J. MacDonald, Y.A. DeClerk, W.E. Laug, Urokinase induces recep-

tor mediated brain tumor cell migration and invasion, J. Neurooncol.

40 (1998) 215–226.

[43] S. Mohanam, R. Sawaya, I. McCutcheon, F. Ali-Osman, D. Boyd,

J.S. Rao, Modulation of in vitro invasion of human glioblastoma cells

by urokinase-type plasminogen activator receptor antibody, Cancer

Res. 53 (1993) 4143–4147.

[44] A. Quattrone, G. Fibbi, E. Anichini, M. Pucci, A. Zamperini, S.

Capaccioli, M.D. Rosso, Reversion of the invasive phenotype of

transformed human fibroblasts by anti-messenger ODNs inhibition

of urokinase receptor gene expression, Cancer Res. 55 (1995) 90–95.

[45] T.A. Libermann, H.R. Nusbaum, N. Razon, R. Kris, I. Lax, H. Soreq,

N. Whittle, M.D. Waterfield, A. Ullrich, J. Schlessinger, Amplifica-

tion, enhanced expression and possible rearrangement of EGF recep-

tor gene in primary human brain tumours of glial origin, Nature 313

(1985) 144–147.

[46] D.E. Neal, C. Marsh, M.K. Bennett, P.D. Abel, R.R. Hall, J.R.C.

Sainsbury, A.L. Harris, Epidermal-growth-factor receptors in human

blood cancer: comparison of invasive and superficial tumours, Lan-

ceti i (1985) 366–368.

[47] J.R.C. Sainsbury, J.R. Farndon, G.K. Needham, A.J. Malcolm,

A.L. Harris, Epidermal-growth-factor receptor status as predictor

of early recurrence of and death from breast cancer, Lanceti

(1987) 1398–1402.

[48] A. Estreicher, J. Muhlhauser, J. Carpentier, L. Orci, J. Vassalli, The

A. Mamoune et al. / Experimental Cell Research 299 (2004) 91–100100

receptor for urokinase type plasminogen activator polarizes expres-

sion of the protease to the leading edge of migrating edge of mono-

cytes and promotes degradation of enzyme inhibitor complexes, J.

Cell Biol. 111 (1990) 783–792.

[49] H. Tang, D.M. Kerins, Q. Hao, T. Inagami, D.E. Vaughan, The uro-

kinase-type plasminogen activator receptor mediates tyrosine phos-

phorylation of focal adhesion proteins and activation of mitogen-

activated protein kinase in cultured endothelial cells, J. Biol. Chem.

273 (1998) 18268–18272.

[50] M. Jo, K.S. Thomas, L. Wu, S.L. Gonias, Soluble urokinase-type

plasminogen activator receptor inhibits cancer cell growth and inva-

sion by direct urokinase-independent effects on cell signaling, J. Biol.

Chem. 278 (2003) 46692–46698.

[51] A. Wells, Molecule in focus: EGF receptor, Int. J. Biochem. Cell Biol.

31 (1999) 637–643.

[52] Y. Okusa, T. Ichikura, H. Mochizuki, N. Shinomiya, Urokinase type

plasminogen activator and its receptor regulates the invasive potential

of gastric cancer cell lines, Int. J. Oncol. 17 (2000) 1001–1005.

[53] Y. Wei, J.A. Eble, Z. Wang, J.A. Kreidberg, H.A. Chapman, Uroki-

nase receptors promote h1 integrin function through interactions with

integrin a3h1, Mol. Biol. Cell 12 (2001) 2975–2986.

[54] T. Moriyama, H. Kataoka, P. Hamasuna, E. Yoshida, T. Sameshima,

T. Iseda, K. Yokogami, S. Nakano, M. Koono, S. Wakisaka, Simul-

taneous up-regulation of urokinase-type plasminogen activator (uPA)

and uPA receptor by hepatocyte growth factor/scatter factor in human

glioma cells, Clin. Exp. Metastasis 17 (1999) 873–879.