| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
INSERM U482, Signal Transduction and Cellular Functions in Diabetes and Digestive Cancers, Hôpital Saint-Antoine, 75571 Paris Cedex 12, France;
* The Laboratory of Experimental Cancerology, Ghent University Hospital, B-9000 Ghent, Belgium;
Novo Nordisk, Bagsvaerd, DK-2880, Denmark;
Department of Pathology, University of Newcastle, Newcastle upon Tyne NE1 4LP, UK; and
INSERM U452, Biologie Cellulaire et Moléculaire des Microorganismes Pathogènes et de leurs Toxines, UFR de Médecine, 06107, Nice Cedex, France
2Correspondence: INSERM Unit U482, Hôpital Saint-Antoine, 75571 Paris Cedex 12, France. E-mail: gespach{at}st-antoine.inserm.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
q (AGq), G12, G13
(AG12/13), which are signaling elements downstream of TXA2-R.
Ectopic overexpression of pS2 cDNA and protein in MDCKts.src-pS2 cells
and human colorectal cancer cells HCT8/S11-pS2 initiate distinct
invasion signals that are Rho independent and COX and TXA2-R dependent.
We detected a marked induction of COX-2 protein and accumulation of the
stable PGH2/TXA2 metabolite TXB2 in the conditioned medium from cells
transformed by src. This led to activation of the TXA2-R-dependent
invasion pathway, which is monitored via a Rho- and
G12/G13-independent mechanism using the Gq/PKC signaling
cascade. These findings identify a new intracrine/paracrine loop that
can be monitored by TFFs and src in inflammatory diseases and
progression of colorectal cancers.—Rodrigues, S., Nguyen, Q.-D.,
Faivre, S., Bruyneel, E., Thim, L., Westley, B., May, F., Flatau, G.,
Mareel, M., Gespach, C., Emami, S. Activation of cellular invasion by
trefoil peptides and src is mediated by cyclooxygenase- and thromboxane
A2 receptor-dependent signaling pathways.
Key Words: inflammation • cancer progression • heterotrimeric G-proteins • Rho-like GTPases • phospholipase C
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4
5)
. The three members of
the trefoil factor family [pS2 (TFF1), spasmolytic polypeptide SP
(TFF2), and intestinal trefoil factor ITF (TFF3)] are expressed in
various tissues including hypothalamus, pituitary, breast, salivary
gland, and prostate (5
, 6)
. TFFs are abundantly cosecreted
and tightly associated with mucins in a region-specific fashion
throughout the gastrointestinal tract and are heat, acid, and protease
resistant (5
, 7
8
9)
. For example, pS2 has shown to be
expressed in mucus-secreting cells of the human stomach
(10)
. Similarly, SP is localized in gastric glands and
duodenal Brunner’s glands, whereas ITF has been identified in goblet
cells of the small intestine and colon (2
, 11)
. In humans,
TFFs are overexpressed or induced in inflammatory conditions such as
peptic ulceration, colitis, Crohn’s disease, pancreatitis, and biliary
disease (5
, 12
, 13)
. TFFs are thought to have beneficial
effects on mucosal integrity via paracrine, local, and distant action
in luminal digestive epithelial cells, after ulceration, and in wound
healing during intestinal damage and inflammatory bowel disease
(14
15
16)
.
Experimental and clinical data strongly suggest a close relationship
between chronic inflammatory diseases and the incidence of solid tumors
in the intestine, breast, and prostate (17
, 18)
.
Inflammatory breast cancer is one of the most aggressive types of human
breast cancer, with a high incidence of local and systemic recurrence
and distant metastases (19)
. Targeted disruption of the
gene encoding the pertussis toxin (PTx) -sensitive heterotrimeric
G-protein subunit Gi2 induces inflammatory colitis that resembles
the human disease in many aspects, including the development of
adenocarcinoma in the colon (20)
. Several clinical studies
have demonstrated that nonsteroidal antiinflammatory drugs (NSAIDs:
sulindac, indomethacin, piroxicam) reduce the number and size of
intestinal polyps in patients with familial adenomatous polyposis
(21
22
23)
and inhibit chemically induced colon
carcinogenesis in rodents (24)
. The common mechanism of
action of NSAIDs is the inhibition of the cyclooxygenase pathway (COX).
Inactivation of the COX-2 gene and selective COX-2 inhibitors provide
protection against the development of intestinal polyps and tumors
(25
, 26)
.
The cyclooxygenase pathway catalyzes the bis-oxygenation of arachidonic
acid into prostaglandin PGH2, the immediate substrate for conversion
into other PGs or thromboxane (TXA2) by specific synthases
(27)
. These COX products are potent inflammatory agents.
For example, TXA2 is suggested as a mediator of acute colitis in the
rat colon. COX-1 is constitutively expressed in nearly all tissues,
whereas COX-2 expression is induced in response to proinflammatory
factors, tumor promoters, oncogenes, and cytokines
(28
29
30
31)
. COX-2 is abnormally elevated in colon adenomas,
colorectal cancers, and azoxymethane-induced mouse tumors
(32
33
34)
.
Inflammatory situations and cancer progression are therefore associated
with up-regulation of TFFs and COX in breast, prostate, and the
gastrointestinal tract. We recently demonstrated intense expression of
pS2 by Western blot and immunohistochemistry in human colorectal tumors
and associated nonmalignant mucosa during the neoplastic progression,
from adenoma through to liver metastases (35)
.
Furthermore, we showed that pS2, SP and ITF induce cellular scattering
and invasion of type I collagen in kidney and colonic epithelial cells,
suggesting that TFFs can be involved in the progression of
gastrointestinal tumors. The first steps in tumor invasion and
metastasis involve destruction of the basement membrane, which contains
type IV collagen, and subsequent degradation of type I collagen, the
major structural component of the extracellular matrix that is degraded
by cancer cells, stroma fibroblasts, and endothelial cells
(36
37
38)
. We have also shown that the small GTPase RhoA,
the src oncogene, and phospholipase C/PKC pathways are essential
signaling elements linking TFFs and cellular invasion
(35)
. Thus, additional adverse effects for TFFs should be
considered in view of their constitutive accumulation in inflamed
digestive tract, human digestive tumors, and their potential actions in
cell migration, scattering, tumor invasion, and metastasis.
The present studies were prompted by our interest in defining the possible functional relations between the cancer-associated signaling pathways using TFFs, src, and COX on the invasive phenotype. Colonic and kidney epithelial cells were treated with exogenous TFFs and the cellular response was compared with stably transfected cells that express pS2 constitutively, reflecting the situation we observed for the pS2 protein in human colorectal cancers. Our hypothesis assumes that cyclooxygenases and their products are involved in the signaling pathways activated by TFFs and src during cellular invasion. In this paper, we demonstrate that the COX and TXA2-R pathways exert a pivotal role in cellular invasion initiated by TFFs and src.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-protein subunit Gq(Q209L) was generously
provided by Dr. E. Peralta (Harvard University, Cambridge, MA). This
sequence was cloned into the EcoRI/XhoI site of
the eukaryotic expression vector pcDNA3.1, which contains the
neoresistance gene (InVitrogen, Groningen, The Netherlands). The
G12(Q229L) and G13(Q226L) sequences encoding the activated
G-proteins G12 (1.8 kb) and G13 (2.4 kb) were inserted into
pcDNA3 at the EcoRI/XbaI and
BamHI/XbaI sites, respectively. These two
pcDNA3/G12Q229L and/G13Q226L vectors were generously provided by
Dr. N. Dhanasekaran (Fels Institute for Cancer Research, Philadelphia,
PA). These mutationally activated forms of GTPase-deficient G-proteins
were designated AGq, AG12, and AG13. The human pS2 cDNA
fragment (462 bp) was excised from the pGEM1 plasmid (39)
by EcoRI/BamHI digestion and cloned in-frame into
pcDNA3. The new plasmid construct was called pcDNA3-hpS2. The structure
and function of pcDNA3-hpS2 was checked by direct DNA sequencing and
expression in HEK-293T cells.
Cell lines and culture conditions
The human embryonic kidney cell line HEK-293T was a generous
gift from Dr. Silvio Gutkind (NIH, Bethesda, MD). Madin-Darby canine
kidney epithelial cells MDCKts.src transformed by a
temperature-sensitive mutant of v-src (MDCKts.src, clone 2) were
cultured in Dulbecco’s modified Eagle’s medium (Gibco BRL, Cergy
Pontoise, France) supplemented with 10% fetal calf serum (Boehringer
Mannheim, Paris, France) plus L-gutamine and antibiotics (Gibco BRL),
as described previously (40)
. MDCKts.src cells display an
invasive phenotype at the permissive temperature of 35°C for src
activity and are not invasive at the restrictive temperature 40°C.
The parental MDCK cell line was a generous gift from Dr. Julian
Downward (Imperial Cancer Research Fund, London, UK). The human
colorectal cell line PCmsrc was grown routinely in 6 cm-diameter Petri
dishes (41)
. After transfer of the activated c-src
oncogene into the premalignant PC/AA/C1 cell line, PCmsrc cells became
tumorigenic in athymic nude mice and invasive upon addition of HGF,
leptin, and TFFs (35
, 41
, 42)
. HCT8/S11 human colonic
cancer cells and MCF-7 human breast cancer cells were cultured as
described (43
, 44)
.
Stable transfection of kidney epithelial cells
About 3 x 106 MDCKts.src cells were
stably transfected with activated forms of the PTx-insensitive
G-protein subunits AGq, AG12, and AG13, using the
corresponding pcDNA3 expression plasmids (3 µg), and 18 µl of
LipofectAMINE Plus reagent (Gibco BRL). Control transfections were
performed using the empty vector pcDNA3. MDCKts.src cells and human
colonic HCT8/S11 cells (also designated HCT-8/E11 cells) were
transfected under the same conditions using the pcDNA3-hpS2 vector
(35)
. After 48 h, cultures were selected for 2 wk in
1 mg/ml neomycin (Gibco BRL); resistant colonies were ring-cloned as
individual colonies or pooled for analysis of ectopic overexpression of
the AGq, AG12, AG13, and pS2 transgenes by immunoblot analysis
and further functional characterization (vide infra). Kidney
MDCKts.src-pS2 cells (clone 2) and human colonic epithelial cells
HCT-8/S11-pS2 cells (clone 2) stably transfected by human full-length
hpS2 were cultured under standard conditions (35
, 44)
.
Conditioned media (CM) from parental MDCKts.src cells and their
counterparts transformed by pS2, src, AGq, AG12, and AG13 were
collected from subconfluent cell layers (70–90% confluence in 3.5 cm
Petri dishes) and used in the invasion and TXB2 assays described below.
The CMs were designated as CM-pS2, CM-src, CM-AGq, CM-A G12, and
CM-AG13, respectively.
Western blot analyses
For immunoblotting, cultured cells were homogenized at 4°C in
RIPA buffer containing 0.1 mg/ml phenylmethysulfonyl fluoride (PMSF),
100 µM benzamidine, and 100 mM
Na3VO4 as protease and
phosphatase inhibitors. Insoluble material was removed by
centrifugation for 15 min at 4°C and 12,000 g. Proteins
were resolved in 12.5% SDS-PAGE gels and transferred to PVDF or
Hybond-C Extra membranes (Amersham Pharmacia Biotech, Orsay, France).
Membranes were blocked overnight in Tris-buffered saline (TBS: 20 mM
Tris-HCl, pH 8, 150 mM NaCl) containing 5% dried skimmed milk. The
blots were then probed for 4 h at room temperature with one of the
following primary rabbit polyclonal antibodies (pAbs) at the indicated
dilutions: the pAb B860–15 specific for G13 (1:1000) was from Dr.
Paul Sternweis (University of Texas, Dallas, TX); the pAbs specific for
G12 (S-20, 1:100), Gq (E-17, 1:200), COX-1 (C-20, 1:500) and the
goat pAb N-20 against COX-2 were from Santa Cruz Biotechnology (TEBU,
Le Perray-en-Yvelines, France). The primary mouse mAbs against
cyclooxygenases COX-2 (1:2000) and COX-1 (1:1000) were from Cayman
Chemical (Spibio, Massy, France). The anti-pS2 mAb p2802 raised against
the carboxyl-terminal domain of pS2 has been described
(10)
. Membranes were washed in TBS containing 0.1% Tween
20 and probed for 90 min with secondary antibodies consisting of a goat
anti-mouse IgGs pAb (1:2000, Santa Cruz Biotechnologies) or a donkey
anti-rabbit IgGs pAb (1:2000, Amersham), then revealed by enhanced
chemiluminescence Western detection (ECL, Amersham).
Collagen invasion assays
For invasion of collagen gels by renal and colorectal epithelial
cells, Petri dishes were filled with 1.35 ml of neutralized type I
collagen and incubated overnight at 37°C to allow gelling. Cells were
harvested and isolated using Moscona buffer and trypsin/EDTA, then
seeded on top of the collagen gels at the density of 0.3 x
106 cells per dish. Dispersed cells
(0.3x106 cells) were cultured for 24 h at
the indicated temperature in the presence or absence of the effectors
and drugs. Invasive and superficial cells were counted in 12 fields of
0.157 mm2. The invasion index is the percentage
of cells invading the gel over the total number of cells
(45)
. None of the agents tested in the 24 h invasion
assay (e.g., wortmannin, NS-398, SC-560, and SQ-295) interfered with
cell growth and viability (trypan blue exclusion test).
Other assays, peptides and reagents
Conditioned culture media were prepared by plating 0.5–1 x 105 cells/well in 6-well culture dishes. Media
from confluent layers was collected and cleared by centrifugation (5000
g, 10 min). Secretion of the cyclooxygenase-derived product
TXA2 in the conditioned culture media was measured indirectly as its
stable thromboxane B2 metabolite, using the EIA enzyme immunoassay kit
from R&D Systems (Abingdon, UK). Intracellular TXB2 was quantified in
the ice-cold cell lysates collected in 250 µl of the EIA buffer after
cellular disruption through a 25 gauge hypodermic needle. Data are
expressed as pg TXB2/µg protein. Clostridium botulinum
exoenzyme C3 transferase (abbreviated as C3T), which ADP-ribosylates
and inactivates the small GTPases RhoA, B, and C but not the ras,
Cdc42, or Rac GTPases (46)
, was isolated at 37°C from
the BL21DE3 strain transformed by the pET28a plasmid recombined with
the C3 gene. The toxin was purified according a modified version of the
method published by Saito and Narumiya (47)
. Briefly,
cells were lysed using a French press and the supernatant was subjected
to anionic exchange chromatography (CM Sepharose fast flow, Amersham
Pharmacia Biotech). Human ITF was produced in yeast and purified as
described (48)
; recombinant hpS2 produced in
Escherichia coli was purified by affinity chromatography
(49)
. TFFs were used as monomers. Hepatocyte growth factor
scatter factor (HGF) was a generous gift from Pr. Paolo Comoglio
(University of Turin, Italy). Leptin was from R&D Systems Europe Ltd.
(Oxon, UK). Wortmannin (WORT, a PI3'-kinase inhibitor), the PLCß
inhibitor U-73122, the COX-1 and COX-2 inhibitors SC-560 and NS-398,
Gö6976 (a selective inhibitor of PKC and PCKß I isoenzymes
that has no effect on the atypical
Ca2+-independent PKCs), and GF109203X
(abbreviated as GF109, a PKC inhibitor with high selectivity for
PKC, ßI, ßII, 
, and 
isoenzymes) were from Calbiochem
(Meudon, France). The prostaglandin H2/thromboxane A2 receptor (TXA2-R)
agonist U-46619 and the TXA2-R antagonist SQ29,548 (designated SQ-295
in the present study) were from Cayman Chemicals. PTx and PMSF were
from Sigma (Saint-Quentin Fallavier, France). Collagen type I was from
Upstate Biotechnology (Lake Placid, NY).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, B
, direct addition of intestinal trefoil factor (ITF, 0.1
µM) induced invasion of collagen gels by kidney epithelial cells
MDCKts.src and human colorectal cancer cells PCmsrc (invasion
index=11±1.7% and 10±0.9%, respectively). Control values
were 1 ± 0.5 and 1 ± 0.6%, respectively. This marked
stimulation of cellular invasion (
10-fold increase) was completely
abolished by the PI3'-K inhibitor wortmannin, the RhoA inhibitor C3T,
the PLC inhibitor U-73122, and the COX-2 inhibitor NS-398 (25
µM).
|
Cellular invasion index measured in MDCKts.src cells was also increased
to 7 ± 1.6% (n=4 experiments) after exogenous
addition of pS2 (Fig. 2A
) and to 11 ± 1.4% after constitutive activation of
src (Fig. 2B
). The COX-1 inhibitor SC-560 blocked the
invasive phenotype induced by pS2 in a dose-dependent manner. It become
apparent that the selective COX-1 and COX-2 inhibitors NS-398 (Fig. 1A
, B
) and SC-560 (Fig. 2A
) abolish cellular
invasion induced by ITF and pS2 in kidney and colonic epithelial cells.
|
Our findings raise the possibility that the cyclooxygenase-derived
products TXA2 and PGH2 are downstream effectors of cellular invasion
induced by TFFs and src. To explore this hypothesis, we first observed
in Fig. 2A
that the TXA2-R agonist U-46619 induced
activation of cellular invasion in MDCKts.src cells (9±1%) that was
comparable to activation by TFFs and src and was abolished by the COX-1
inhibitor SC-560. Second, we found that cellular invasion induced by
src in MDCKts.src cells was reversed by NS-398 and SC-560 (Fig. 2B
). Third, the PI3'-K, Rho, and PLC inhibitors wortmannin,
C3T, and U-73122 abolished src-induced cellular invasion in MDCKts.src
cells (Fig. 2B
). Most important, the TXA2-R antagonist
SQ-295 also blocked cellular invasion induced by the src oncogene.
Therefore, we sought to extend these results to trefoil factors pS2 and
ITF.
As shown in Fig. 3
, the COX-2 inhibitor NS-398 prevented the induction of invasion by pS2
and ITF (Fig. 3A
, B
), but was ineffective against
HGF-induced invasion in MDCKts.src cells (Fig. 3C
) or leptin
(Fig. 5B
, vide infra). Similarly, the TXA2-R antagonist
SQ-295 (10 and 30 µM) failed to control invasion induced by HGF and
leptin (data not shown), indicating that COX and the TXA2-R selectively
play an essential role in TFF- and src-induced invasion. This implies
that these two COX/TXA2-R pathways are not involved in this cellular
activity induced by the leptin receptor Ob-Rb or the HGF receptor Met
(42)
. In contrast, the TXA2-R antagonist SQ-295
dose-dependently reversed pS2 and ITF signaling (Fig. 3A
, B
), whereas the TXA2-R mimetic U-46619 dose-dependently induced
invasiveness (Fig. 3C
). An apparent maximal effect (invasion
index=10±0.8%) was induced at 10 µM U-46619. However, it is
noteworthy that invasiveness induced by this concentration of U-46619
was not inhibited by C3T, indicating that an additional invasion
pathway, independent of Rho, is operative. This is addressed in more
detail in the experiments described below.
|
|
Cellular invasion induced by ectopic pS2 overexpression is
dependent on PLC, COX-1/COX-2, but is Rho independent
As shown in Fig. 4A
, B
), MDCKts.src-pS2 cells (clone 2) and human colonic
epithelial cells HCT8/S11-pS2 (clone 2) stably transfected by the human
pS2 cDNA exhibited constitutive invasiveness (10±0.5 and 11±0.3%)
that was abrogated by wortmannin, U-73122, and the COX-2 inhibitor
NS-398. Abrogation of invasiveness by HCT8/S11-pS2 cells was also found
in the presence of the COX-1 inhibitor SC-560 (not shown). In contrast,
the Rho inhibitor C3T failed to reverse invasion induced by the ectopic
expression of pS2. Overexpression of pS2 was confirmed by
immunoblotting in the clonal derivatives isolated from MDCKts.src-pS2
cells (Fig. 4C
: pool 1 and clones 2, 5) and HCT8/S11-pS2
cells (Fig. 4D
: pool 1 and clones 2, 6). The same signal was
identified in two independent positive controls, using a human gastric
tumor sample and the MCF-7 human breast cancer cell line. We next
tested whether we could detect the pS2 protein in the CM from kidney
and colonic epithelial cells stably transfected by the pS2 cDNA. The
presence of the main 9.6 kDa band and a diffuse band with higher
molecular weight was evident in the CM-pS2 isolated from HCT8/S11-pS2
and MDCKts.src-pS2 cells, but was absent in their nontransfected
counterparts (Fig. 4E
). Specificity of the detected
pS2-immunoreactive band (9.6 kDa) was established using transient
transfections in the human embryonic kidney cell line HEK-293T (Fig. 4F
).
|
Induction of cellular invasion by the conditioned media from
MDCKts.src-pS2 cells stably transfected by the human pS2 cDNA
We next assayed the biological activity of secreted products in
the conditioned medium (CM-pS2) prepared from MDCKts.src-pS2 cells in
our invasion assay, using MDCKts.src cells and parental MDCK cells as
targets (Fig. 5A
, B
). As shown in Fig. 5A
, CM-pS2 induced a
remarkable activation of invasiveness in MDCKts.src cells (invasion
index=10.3±1.3%) that was blocked by wortmannin and the PLC inhibitor
U-73122. Invasion induced by CM-pS2 was resistant to C3T and PTx,
suggesting that the Rho-dependent pathway and PTx-sensitive
heterotrimeric proteins (Go/Gi1–3) are not involved. The CM from
MDCKts.src-pcDNA3 cells used as a negative control was ineffective.
Furthermore, the effect of CM-pS2 was not inhibited by C3T, strongly
suggesting that the invasiveness of MDCKts.src challenged with CM-pS2
is not related to the biological activity of pS2 in the CM-pS2. To
investigate this, the effect of the CM-pS2 on invasiveness of the
parental MDCK cells that are not sensitive to ITF (35)
or
pS2 (Fig. 5B
) was examined. CM-pS2 markedly enhanced
invasiveness in nontransformed MDCK cells (invasion index=10±0.5%).
This induction by CM-pS2 was completely blocked by the TXA2-R
antagonist SQ-295. As controls, we verified that stimulation of MDCK
cells by the TXA2-R agonist U-46619 led to efficient stimulation,
whereas the COX-2 inhibitor NS-398 was ineffective against leptin
receptor activation.
Activation of cellular invasion by the GTPase-deficient forms of
the PTx-insensitive G-proteins Gq, G12, and G13
Having established the causal role of the COX and TXA2-R pathways
in cellular invasion stimulated by TFFs and src, we next investigated
how TXA2-R might stimulate the ability of parental MDCK cells and
MDCKts.src cells to invade collagen gels. The cyclooxygenase-derived
products TXA2 and PGH2 are potent activators of the TXA2-R. Various
authors have proposed that the G-protein-coupled Tp and Tpß
receptor subtypes for thromboxane A2 and PGH2 (designated
here as TXA2-R) might couple to several G-proteins, including the
PTx-insensitive G-proteins Gq, G12/G13, G16, and to the
PTx-sensitive Gi2 (50
51
52
53)
. The data presented in Fig. 5A
support the notion that the PTx-sensitive G-protein
Gi2 is not involved. The major signaling pathway used by both
Tp/ß subtypes of the TXA2-R is the Gq-dependent stimulation of
the ß-isoforms of phospholipase C (PLCß) by TXA2/PGH2 agonists,
resulting in protein kinase C activation (PKC) and mobilization of
intracellular calcium. It is therefore plausible that TFFs and src
promote invasiveness through activation of the TXA2-R signaling pathway
using the PTx-insensitive G-proteins Gq, G12, and G13. We
therefore extended our studies to investigate the effects of the
GTPase-deficient mutants of Gq, G12, and G13 encoding
activated forms (AGq, AG12, AG13) of these three G-protein
subunits. Immunoblot analysis identified several positive MDCKts.src
clones expressing AGq (clones 2, 3, and 9), AG12 (clones 1 and
4), and AG13 (clones 2 and 3). Densitometric analysis revealed that
these ectopic G-protein mutants were overexpressed two- to threefold
over endogenous levels (data not shown). Stably transfected
MDCKts.src-AGq cells (clone 9) exhibited constitutive invasiveness
(invasion index=9±1.3%) that was resistant to the Rho inhibitor C3T
(Fig. 6A
). All the elements related to the TFF and src signaling
pathways, including PI3'-K, PLC, COX-1, COX-2, and the TXA2-R, appear
to be essential components for cellular invasion induced by the
G-protein AGq (Fig. 6A
). As expected, the PKC inhibitors
Gö6976, GF109, and down-regulation of PKC by phorbol ester
abolished invasion caused by AGq in MDCKts.src cells (not shown).
The activated forms of AG12 and AG13 (Fig. 6B
, C
) also
induced activation of invasiveness in MDCKts.src cells (invasion
index=10±1.7 and 10±1%), but showed a different pattern of
inhibition. Whereas U-73122, NS-398, and SQ-295 had no effect on
AG12/AG13-induced invasion, the Rho and PI3'K inhibitors
completely reversed this induction. This last result is consistent with
the identification of RhoA as a downstream target through which
activated G12 and G13 induce cytoskeletal reorganization.
Finally, it is now well accepted that the ubiquitously expressed
G12/G13 families of heterotrimeric G-proteins have been
implicated in cell transformation (54
, 55)
.
|
Taken together, our data support the notion that Gq and its
downstream targets, including PLC-ß/PKC, play an essential role in
the induction and maintenance of the invasive phenotype in kidney and
colonic epithelial cells stimulated by TFFs, src, and the TXA2-R. To
confirm this hypothesis, we compared the ability of conditioned medium
from AGq, AG12, and AG13-transfected MDCKts.src cell cultures
to initiate invasiveness in the parental MDCK cell line. As shown in
Fig. 7A
, CM-Gq mimicked the invasion induced by ectopic pS2 in
MDCKts.src cells and HCT8/S11 cells (Fig. 4A
, B
),
because the same Rho-independent signaling cascades using PI3'-K, PLC
and COX are involved. In contrast, CM from MDCKts.src cells transformed
by AG13 or AG12 (CM-AG12 or -AG13) did not induce
invasiveness (Fig. 7B
).
|
Similar to the response observed with CM-Gq, conditioned
medium isolated from the src-transformed MDCKts.src cells cultured at
the permissive temperature 35°C (CM-src) led to marked invasiveness
in parental MDCK cells (invasion index=11%), as compared to leptin as
positive control (Fig. 7C
). Consistent with these findings,
invasion induced by CM-src was also blocked by pharmacological
inhibitors of the PLC and COX pathways, but not by the Rho inhibitor
C3T.
Expression of cyclooxygenases and TXB2 in kidney and colonic
epithelial cells transformed by pS2, src, and AGq
As shown in Fig. 8A
, immunoreactive COX-2 protein expression in MDCKts.src
cells was markedly induced by activated src and AGq. Elevation of
COX-2 levels by src was not blocked by C3T, U73122, or the
cyclooxygenase inhibitors NS398 and SC-560, but was reversed by
wortmannin, suggesting that PI3'-K is implicated in src-induced COX-2
expression. To ascertain the functionality of immunoreactive COX-2
induction, accumulation of the stable PGH2/TXA2 metabolite TXB2 was
measured in CM-src. Activation of src resulted in a remarkable increase
of TXB2 release in CM of MDCKts.src cells (33-fold control), from a
basal level of 0.44 ± 0.22 to 14.8 ± 2.9 pg/µg protein
(n=7 experiments). There was no elevation of TXB2 levels in
response to HGF, leptin, and AGq. Accumulation of TXB2 induced by
src in CM-src was blocked by wortmannin, attenuated by the PLC
inhibitor U73122 (P<0.02), and was C3T-insensitive (data
not shown). Src-induced TXB2 levels were abolished by SC-560 and
NS-398, suggesting that the functional activities of COX-1 and COX-2
are interdependent (Fig. 8B
). Half-maximal inhibition by
SC-560 was observed at 3 nM for invasion and TXB2 accumulation.
|
Regarding the COX-2 inhibitor, half-maximal inhibition of src-induced
invasion and TXB2 accumulation was observed at 2–5 nM NS-398 in the
endogenous COX-2/COX-1 system presented here. Similarly, intracellular
TXB2 levels induced by src in MDCKts.src cells were dose-dependently
inhibited by the COX inhibitors, with inhibitory potencies of 5 nM
SC-560 and 1 nM NS-398 (not shown). Whereas overexpression of pS2 led
to a much lower extent of COX-2 protein induction in MDCKts.src-pS2
cells (Fig. 8A
) and only a twofold increase of TXB2 levels
in MDCKts.src cells and HCT8/S11 cells (basal level=0.039±0.02 pg/µg
protein, n=7 experiments), cellular invasion induced by pS2
in MDCKts.src-pS2 cells (ectopic overexpression) and HCT8/S11 cells
(external addition) was abrogated by 100 nM SC-560 or NS-398
(IC50=3–5 nM SC-560 and 1–4 nM NS-398).
Induction of COX-2 protein in HCT8/S11-pS2 cells and MDCKts.src-AGq
cells (Fig. 8A
) was insensitive to C3T, U-73122, COX-1/COX-2
inhibitors, and wortmannin.
It appears that depletion of TXB2 and blockade of COX-1/COX-2 enzymes
by NS-398 and SC-560 may prove to be causal in the abolition of
invasiveness induced by src and TFFs in kidney and colonic cancer
cells. The cyclooxygenase COX-1 is constitutively expressed in most
cell types, whereas the immediate-early response gene COX-2 is induced
by v-src, ras, lipopolysaccharide (LPS), cytokines, and tumor
promoters, such as phorbol esters or bile acids (28
29
30
31
, 56
57
58)
. Accordingly, when MDCKts.src or HCT8/S11 cells are
transformed by src, pS2, or Gq, COX-1 protein levels are unchanged
(Fig. 8A
) even in the presence of pharmacological inhibitors
of the src- and TFF-dependent invasion pathways. Ectopic overexpression
of the COX-1 or COX-2 cDNAs also resulted in the activation of membrane
metalloproteinase MMP-2, tumorigenesis, and metastatic potential
(59
, 60)
.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Because the TXA2-R and its associated G-proteins Gq, G12, and
G13 are candidate effectors linking COXs and cancer progression, we
have demonstrated that COX-2 and TXB2 levels are remarkably elevated in
MDCKts.src cells during src-induced invasiveness. Thus, the src
oncogene can generate, first, a COX/TXA2-dependent intracrine/autocrine
activation loop via the TXA2-R/Gq pathway, and second, a paracrine
activation of cellular invasion via increased levels of TXA2 in the
CM-src. Downstream activation of the TXA2-R associated Gq subunit is
directly connected with the PLC pathway and its downstream elements,
including PKCs and calcium-dependent signaling. There are several lines
of evidence that these routes are indeed involved in cellular invasion
and metastasis. Accordingly, invasion induced by G12/G13 is Rho
dependent and unaffected by PLC and COX inhibitors. In addition, the CM
from G12/G13-transfected MDCKts.src cells does not induce
invasion. Most important, the Rho inhibitor C3T failed to abolish
invasion induced by ectopic overexpression of pS2 and AGq,
suggesting that these two signaling pathways may overcome and
circumvent this blockade.
|
Although src activation initiated a remarkable increase in TXB2 and
COX-2 protein levels in MDCKts.src cells, this cyclooxygenase product
is not increased consistently in the conditioned medium from pS2- and
AGq-transformed kidney and colonic epithelial cells. We therefore
performed additional experiments showing that the dose-response of
NS-398 and SC-560 for inhibition of invasion and TXB2 accumulation
induced by src in MDCKts.src cells are superimposed, according to
IC50 values that are consistent with their
respective potencies on the COX-1 and COX-2 enzymes. Regarding the
COX-1 inhibitor, half-maximal inhibition by SC-560 was observed at 3 nM
for invasion and TXB2 accumulation. The IC50 of
SC-560 on recombinant COX-1 protein is 10 nM, and SC-560 inhibits
rCOX-2 at 600-fold higher concentration (61)
. Regarding
the COX-2 inhibitor, half-maximal inhibition of src-induced invasion
and TXB2 accumulation was observed in the interdependent COX-2/COX-1
system presented here at 3–5 nM NS-398. The potency of NS-398 on
baculovirus-purified COX-2 is 100 nM (62)
. Furthermore, we
found that cellular invasion induced by pS2 in MDCKts.src cells and
HCT8/S11 cells was abolished by the COX-1/COX-2 inhibitors within the
same range of concentrations. Therefore, our data provide a direct
functional link between cyclooxygenases and cellular invasion induced
by src and TFFs, and underscore that invasiveness is abrogated when
TXB2 levels and cyclooxygenase activities are fully down-regulated.
This relationship strongly suggests that transient, local, and marginal
elevation of TXA2 levels that are not associated with a detectable
increase of cellular and external TXB2 in the CM can exert a crucial
role in the activation of cellular invasion and the TXA2-R/Gq
signaling system via local diffusion within the plasma membrane.
In vitro, the biological half-life of TXA2 is 30 s. Because PGH2 is
a freely diffusible molecule in the membrane lipid bilayer, one can
speculate that this TXA2-R ligand can be rapidly withdrawn from the
external milieu after interaction with this membrane receptor.
Diffusion of endogenous PGH/TXA2 products from discrete
intracellular/intramembrane pools, leading to intracrine- and
autocrine-dependent activation of cellular invasion is also consistent
with our data. COX-2 protein is rapidly degraded via the proteasome
pathway in human colon carcinoma cells, which constitutively express
COX-2 (63)
. A recent report by Coffey and co-workers
(64)
demonstrated that EGF-R activation leads to the
basolateral release of prostaglandins in intestinal epithelial cells.
Such a polarized secretory pathway for COX-produced eicosanoids can be
related to the short lifetime of TXA2 and COX-2, linked to a
compensatory local delivery, at the immediate vicinity of this
transmembrane receptor. In favor with this proposition, recent reports
identified COX-2 expression in both the epithelial and stromal
compartments of sporadic human colorectal cancers and prostate
adenocarcinomas (65
66
67)
. Administration of the
proinflammatory agent LPS induced COX-2 expression in subepithelial
fibroblasts, which are immediately adjacent to intestinal epithelial
cells (68)
. Consistent localization of COX-2 was also
demonstrated in superficial interstitial macrophages in human sporadic
colorectal adenomas (65)
. Therefore, exogenous TXA2 and
PGH2 produced by the tumor stroma, including mesenchymal, endothelial,
immune cells, and platelets, may participate in paracrine activation of
the TXA2-R signaling pathways in cancer cells.
Our findings also imply that the TXA2-R and Gq-dependent activation
mechanism presented in Fig. 9
is controlled by the COX pathways by
preferential coupling of enzyme activity to discrete
prostanoid/thromboxane/prostaglandin synthases, according to the
transformation phenotype induced by TFFs and the src oncogene. Indeed,
we have been unable to detect any elevation of TXB2 levels in
MDCKts.src-AGq cells despite the marked induction of COX-2 protein
by this constitutively activated G-protein. In fact, other products of
the COX pathways that are natural ligands of the TXA2-R (such as PGH2,
other prostanoids and isoprostanes) are potential effectors for this
activation loop. The COX-1/COX-2-derived prostane isomers 8-epi-PGE2
and 8-epi-PGF2 (69
, 70)
whose
effects can be prevented by TXA2-R antagonists are also induced by
chronic cigarette smoking, oxidative stress, and reactive oxygen
species. Alternatively, our data also argue that the functional
competence of the TXA2-R exerts a permissive synergism on invasion
promoters and signaling pathways. As previously observed in
MDCKts.src-transformed cells and human breast cancer MCF-7 cells, the
CM-src in our studies also contains other invasion promoters, such as
E-cadherin fragments (71)
. Similarly, pro-EGF has shown to
be activated by proteases via a G-protein receptor activation loop
similar to the TXA2-R/Gq cascade in the present study
(72)
.
Several studies have addressed the role of cyclooxygenases in
neoplasia, but the molecular mechanisms involved are still poorly
understood. The data reported here indicate that the COX and TXA2-R
signaling pathways may participate to the persistent activation of
Gq linked to the acquisition of the invasive phenotype controlled by
TFFs and src in transformed kidney and colonic epithelial
cells. We propose that this cyclooxygenase- and TXA2-R-dependent
mechanism is exacerbated by the src oncogene and is operative in
TFF-transformed cells (Fig. 9)
. This intracrine and paracrine
activation loop constitutes a potential link for our understanding of
the molecular and cellular mechanisms involved in the chemoprevention
of colorectal cancers and inflammation by NSAIDs. There is a growing
interest in developing NSAID-like drugs for prevention of many types of
human cancers. Because of the gastrointestinal toxicity and frequency
of complications associated with NSAIDs, one can envisage that the
TXA2-R signaling cascade might represent a new therapeutic target with
which to control the multistep progression induced by the
cyclooxygenase pathways in solid tumors.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication November 27, 2000.
Revision received March 7, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
i2-deficient mice. Nature Genetics (London) 10,143-150[Medline]
716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell 87,803-809[Medline]
-catenin in human colon cancer cells. Cancer Res 55,4722-4728 subunit of G12 potentiates the eicosanoid pathway and is highly oncogenic in NIH 3T3 cells Proc. Natl. Acad. Sci. USA 90,6741-6745 subunits of G12 and G13 proteins induce neoplastic transformation of Rat-1 fibroblasts. Oncogene 9,2559-2565[Medline]
by human monocytes. Discriminate production by reactive oxygen species and prostaglandin endoperoxide synthase-2. J. Biol. Chem. 271,8919-8924This article has been cited by other articles:
|
|
 |
|
  N. Amiry, X. Kong, N. Muniraj, N. Kannan, P. M. Grandison, J. Lin, Y. Yang, C. M. Vouyovitch, S. Borges, J. K. Perry, et al. Trefoil Factor-1 (TFF1) Enhances Oncogenicity of Mammary Carcinoma Cells Endocrinology, October 1, 2009; 150(10): 4473 - 4483. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. LeBrun, T. Baetz, C. Foster, P. Farmer, R. Sidhu, H. Guo, K. Harrison, R. Somogyi, L. D. Greller, and H. Feilotter Predicting Outcome in Follicular Lymphoma by Using Interactive Gene Pairs Clin. Cancer Res., January 15, 2008; 14(2): 478 - 487. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. P. Prevost, M. O. Lonchampt, S. Holbeck, S. Attoub, D. Zaharevitz, M. Alley, J. Wright, M. C. Brezak, H. Coulomb, A. Savola, et al. Anticancer Activity of BIM-46174, a New Inhibitor of the Heterotrimeric G{alpha}/G{beta}{gamma} Protein Complex. Cancer Res., September 15, 2006; 66(18): 9227 - 9234. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Jain, G. Chakraborty, and G. C. Kundu The Crucial Role of Cyclooxygenase-2 in Osteopontin-Induced Protein Kinase C {alpha}/c-Src/I{kappa}B Kinase {alpha}/{beta}-Dependent Prostate Tumor Progression and Angiogenesis. Cancer Res., July 1, 2006; 66(13): 6638 - 6648. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C L Wilson, A H Sims, A Howell, C J Miller, and R B Clarke Effects of oestrogen on gene expression in epithelium and stroma of normal human breast tissue. Endocr. Relat. Cancer, June 1, 2006; 13(2): 617 - 628. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Smid, Y. Wang, J. G.M. Klijn, A. M. Sieuwerts, Y. Zhang, D. Atkins, J. W.M. Martens, and J. A. Foekens Genes Associated With Breast Cancer Metastatic to Bone J. Clin. Oncol., May 20, 2006; 24(15): 2261 - 2267. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Michel, J.-S. Silvestre, L. Waeckel, S. Corda, T. Verbeuren, J. P. Vilaine, M. Clergue, M. Duriez, and B. I. Levy Thromboxane A2/Prostaglandin H2 Receptor Activation Mediates Angiotensin II-Induced Postischemic Neovascularization Arterioscler Thromb Vasc Biol, March 1, 2006; 26(3): 488 - 493. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. M. Vestergaard, M. Borre, S. S. Poulsen, E. Nexo, and N. Torring Plasma Levels of Trefoil Factors are Increased in Patients with Advanced Prostate Cancer Clin. Cancer Res., February 1, 2006; 12(3): 807 - 812. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. G. Bottone Jr., Y. Moon, J. S. Kim, B. Alston-Mills, M. Ishibashi, and T. E. Eling The anti-invasive activity of cyclooxygenase inhibitors is regulated by the transcription factor ATF3 (activating transcription factor 3) Mol. Cancer Ther., May 1, 2005; 4(5): 693 - 703. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Liu, A. Stromberg, H.-H. Tai, and J. A. Moscow Thiamine Transporter Gene Expression and Exogenous Thiamine Modulate the Expression of Genes Involved in Drug and Prostaglandin Metabolism in Breast Cancer Cells Mol. Cancer Res., August 1, 2004; 2(8): 477 - 487. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. W. Ashton, Y. Cheng, A. Helisch, and J. A. Ware Thromboxane A2 Receptor Agonists Antagonize the Proangiogenic Effects of Fibroblast Growth Factor-2: Role of Receptor Internalization, Thrombospondin-1, and {alpha}v{beta}3 Circ. Res., April 2, 2004; 94(6): 735 - 742. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Mareel and A. Leroy Clinical, Cellular, and Molecular Aspects of Cancer Invasion Physiol Rev, April 1, 2003; 83(2): 337 - 376. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. RODRIGUES, E. VAN AKEN, S. VAN BOCXLAER, S. ATTOUB, Q.-D. NGUYEN, E. BRUYNEEL, B. R. WESTLEY, F. E. B. MAY, L. THIM, M. MAREEL, et al. Trefoil peptides as proangiogenic factors in vivo and in vitro: implication of cyclooxygenase-2 and EGF receptor signaling FASEB J, January 1, 2003; 17(1): 7 - 16. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Attoub, C. Rivat, S. Rodrigues, S. Van Bocxlaer, M. Bedin, E. Bruyneel, C. Louvet, M. Kornprobst, T. Andre, M. Mareel, et al. The c-kit Tyrosine Kinase Inhibitor STI571 for Colorectal Cancer Therapy Cancer Res., September 1, 2002; 62(17): 4879 - 4883. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Q.-D. NGUYEN, S. FAIVRE, E. BRUYNEEL, C. RIVAT, M. SETO, T. ENDO, M. MAREEL, S. EMAMI, and C. GESPACH RhoA- and RhoD-dependent regulatory switch of G{alpha} subunit signaling by PAR-1 receptors in cellular invasion FASEB J, April 1, 2002; 16(6): 565 - 576. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
