Institut National de la Santé et de la Recherche
Médicale U482, Hôpital Saint-Antoine, Paris, France (S.F.,
K.R., Q.-D.N., S.E., C.G.); and The Laboratory of Experimental
Cancerology, Ghent University, Ghent, Belgium (E.B., M.M.)
It was shown previously that platelet-activating factor receptors
(PAF-Rs) inhibit invasiveness of colonic and kidney epithelial cells
induced by the src and Met oncogenes via a pertussis toxin-sensitive mechanism. Therefore, Madin-Darby canine kidney (MDCKts.src) cells were
stably transfected with constitutively activated forms of G
o,
G
i1, G
i2, G
i3 (AG
o/i), two G
sequestering proteins [C-terminal end of
-adrenergic receptor kinase (ct-
ARK) and the
G
t subunit of retinal G-protein transducin], and G
1-G
2 subunits alone or in combination. Cellular invasion induced by src,
Met, and leptin was abrogated by the AG
o/i, ct-
ARK, and G
t-positive clones, but was induced by coexpression of G
1
2. In
contrast, invasion stimulated by the trefoil factors (TFFs) pS2 and
intestinal trefoil factor in MDCKts.src cells or human colonic
epithelial cells PCmsrc and HCT8/S11 was insensitive to PAF, AG
o,
AG
i1, and AG
i2, but was abolished by AG
i3 and the protease-activated receptor-1 (PAR-1) agonist thrombin
receptor-activating peptide. Depletion of free G
heterodimers by ct-
ARK resulted in a remarkable decrease of cellular
adhesion and spreading on collagen matrix. Our data demonstrate the
following: 1) PAF-Rs impair cellular invasion induced by src, Met, and
leptin via the activation of G
o and G
i1 to -3; 2) invasion
induced by TFFs is selectively inhibited by PAR-1 receptors and G
i3
activation; and 3) G
dimers are required as positive effectors of
invasion pathways induced by oncogenes and epigenetic factors. Thus,
redistribution of G
o/G
i and G
/
heterotrimeric G-proteins by
PAF-R and PAR-1 exert differential functions on positive and negative
signaling pathways involved in cellular invasion and may serve as
potential targets for anticancer therapy.
 |
Introduction |
Persistent
dysregulation of signal transduction pathways promote the acquisition
of anarchic functions in digestive epithelial cells, adjacent tissues,
and immune cells, leading to local inflammation and further
susceptibility to the neoplastic progression. Strong experimental and
clinical data suggest a close relationship between inflammatory disease
and the occurrence of solid tumors in intestine, breast, and prostate
(Rudolph et al., 1995a
; Emami et al., 2001
). Local mediators,
cytokines, and inflammatory agents, such as platelet-activating factor
(PAF), thrombin, and trefoil factors (TFFs) are involved in the
cellular responses to injury and wound repair during human inflammatory
processes in the gastrointestinal tract, including peptic ulceration
and colitis, Crohn's disease, pancreatitis, and biliary disease
(Cirino et al., 1996
; Kotelevets et al., 1998
; Emami et al., 2001
). We
have shown previously that functional and specific PAF receptors
(PAF-R) are expressed in normal human colonic epithelial crypts,
colonic adenoma, and adenocarcinoma cell lines (Kotelevets et al.,
1998
). Moreover, we found that src- and Met-induced invasion of
collagen gels was abrogated by PAF-R activation in Madin-Darbin canine
kidney (MDCKts.src) and colonic PCmsrc cells via a pertussis toxin
(PTx)-sensitive pathway, suggesting that G
o/G
i and G
/
heterotrimeric G-proteins are involved in this process. PTx, which is
specific for G
o and G
i subunits, prevents the catalysis of
GDP-GTP exchange by the receptor, blocking both G
and G
signaling. The preferred substrates of PTx are G
o/i subunits
associated with G
complexes, whereas monomeric G
subunits are
very poor substrates (Rudolph et al., 1995b
). Redistribution and
functional activity of G
subunits released from PTx-sensitive
proteins and serpentine receptors are important components of
exocytosis and scaffold of molecular complexes at the plasma membrane
level and subcellular compartments, including the actin network (Hamm,
1998
). We also presented evidence that leptin and the TFFs pS2,
spasmolytic polypeptide, and intestinal trefoil factor (ITF) are
positive and potent effectors of cellular invasion in premalignant and
src-transformed kidney and colonic epithelial cells (Attoub et al.,
2000
; Emami et al., 2001
). Leptin and TFFs are expressed in epithelial
cells along the gastrointestinal tract and are considered local and
paracrine regulators of mucosal integrity, renewal, inflammation, and
neoplastic progression.
This study was therefore conducted to identify the subtypes and
cellular functions of the heterotrimeric G-protein subunits linked to
the negative control exerted by the serpentine PAF-R on positive
invasion pathways controlled by src, Met, and the intestinal mediators
leptin and TFFs. For this purpose, we stably transfected MDCKts.src
cells by constitutively activated GTPase-deficient forms of G
o,
G
i1, G
i2, G
i3 (AG
o/i), the chimeric C-terminal end
of the
-adrenergic receptor kinase (ct-
ARK) scavenging G
subunits (Pitcher et al., 1992
; Crespo et al., 1994
), and transducin G
subunit (G
t), another agent known to sequester free G-protein 
dimers (Federman et al., 1992
). To examine further the role of
G
subunits on cellular invasion, G
1 was also overexpressed, either alone or with G
2. The relative contribution of PAF-R, AG
o/i, and G
subunits on cellular invasion and adhesion was then examined with the use of these experimental models, as well as
with the human colorectal cell lines PCmsrc and HCT8/S11, which are
derived from familial and sporadic tumors, respectively (Behrens et
al., 1993
; Vermeulen et al., 1995
; Empereur et al., 1997
). Because
trefoil peptides are overexpressed in digestive epithelial cells during
inflammatory situations and cancer progression (Emami et al., 2001
), we
next investigated whether PAF-R can control cellular invasion in
MDCKts.src and HCT8/S11 cells stably transfected by the human pS2 cDNA.
Results were compared with the functional activity of the thrombin
PAR-1 receptor that is specifically activated by the thrombin
receptor-activating peptide (TRAP) (Seiler et al., 1996
). The PAR-1
receptor has multiple-signaling capacity and is also coupled to the
PTx-sensitive G-proteins G
o, G
i1, G
i2, and G
i3. We report
that activated G
o/i subunits exert an invasion-suppressor role,
whereas the G
subunits are required as critical and positive
mediators of cellular invasion pathways induced by oncogenes and
epigenetic factors.
 |
Materials and Methods |
DNA Constructs.
The rat cDNAs encoding the constitutively
activated forms of the PTx-sensitive G
protein subunits AG
o/i
were generously provided by Dr. E. Peralta (Harvard University,
Cambridge, MA). They were cloned in frame into the eukaryotic
expression vector pcDNA3.1 recombined with the neo resistance gene
(Invitrogen BV, Breda, The Netherlands). The AG
o/i cDNAs were
inserted at the cloning sites EcoRI/XhoI
(Q205L-G
o) or EcoRI/XbaI
(Q204L-G
i1, -2, and -3). These mutationally
activated forms of GTPase-deficient G-proteins were further designated
as AG
o/i1, -2, and -3. The structure and function of the expression
vectors encoding AG
o/i was checked by direct DNA sequencing. The
capture and sequestration of G
subunits was induced by the
myc-tagged expression vector pcDNA3 encoding a chimeric protein
comprising the membrane-bound CD8 receptor and ct-
ARK, as described
previously (Pitcher et al., 1992
). The human transducin G
t subunit
cDNA cloned into the pcDNA3.1 expression vector was purchased from the
American Type Culture Collection (Manassas, VA). The bovine G
1 and
G
2 cDNAs were cloned from the pcDM8-1 vector into the pcDNA3.1
expression vector (Crespo et al., 1994
). The G
sequestrant vector
ct-
ARK and expression vectors encoding G
1 and G
2 were
generously provided by Dr. R. Weitzker (Klinikum der
Friedrich-Schiller-Universität, Jena, Germany).
Cell Lines and Culture Conditions.
MDCK 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 (Invitrogen SARL, Cergy Pontoise, France) supplemented
with 10% fetal calf serum (Roche Molecular Biochemicals, Meylan,
France) plus L-glutamine and antibiotics (Invitrogen), as described previously (Behrens et al., 1993
). 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 human colorectal cell line PCmsrc was grown in 6-cm diameter Petri dishes. After transfer of the activated c-src oncogene in the premalignant PC/AA/C1 cell line, PCmsrc
cells became tumorigenic in the athymic nude mice and are invasive upon
addition of hepatocyte growth factor (HGF) (Empereur et al., 1997
). The
MDCKp110* cell line stably transfected with a constitutively activated
form of bovine p110*
by addition of the C-terminal farnesylation
signal from Ha-Ras (Khwaja et al., 1997
) was a generous gift from Dr.
J. Downward (Imperial Cancer Research Fund, London, UK). Kidney
MDCKts.src-pS2 cells (clone 2) and colonic epithelial cells
HCT-8/S11-pS2 cells (clone 2) stably transfected by the human
full-length hpS2 were cultured under standard conditions (Emami et al.,
2001
).
Stable Transfection of Kidney Epithelial Cells.
Approximately 3 × 106 MDCKts.src cells were
stably transfected by the activated forms of the PTx-sensitive
G
-protein subunits AG
o/i1, -2, and -3 using the corresponding
pcDNA3.1 expression plasmids (3 µg) and 18 µl of the LipofectAMINE
Plus reagent (Invitrogen). Control transfections were performed
using the empty vector pcDNA3.1. MDCKts.src cells were also transfected
under the same conditions using either the G
sequestrant vectors
CD8-ct-
ARK or G
t, and the vectors encoding bovine G
1
and G
2 (see above). After 48 h, cultures were selected for 2 weeks in 1 mg/ml neomycin or 0.5 mg/ml hygromycin (Invitrogen) and
MDCKts.src-resistant colonies were ring-cloned as individual colonies
or pooled for analysis of ectopic expression of the AG
o/i,
CD8-ct-
ARK, G
t, G
1, and G
2 proteins by immunoblot
analysis, indirect immunocytochemistry, and additional functional characterization.
Western Blot Analyses, Immunocytochemistry, and Reverse
Transcriptase-Polymerase Chain Reaction (RT-PCR) Analysis.
For
immunoblotting, cultured cells were homogenized at 4°C in
radioimmunoprecipitation buffer containing 0.1 mg/ml
phenylmethylsulfonyl fluoride, 100 µM benzamidine, and 100 µM
Na3VO4 as protease inhibitors. Insoluble
material was removed by centrifugation for 15 min at 4°C and
12,000g. Proteins were resolved using nonreducing
conditions in 12.5% SDS-polyacrylamide gel electrophoresis and
transferred to polyvinylidene difluoride or Hybond-C Extra membranes
(Amersham Pharmacia Biotech, Orsay, France). Membranes were blocked
overnight in Tris-buffered saline (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
polyclonal or monoclonal antibodies at the indicated dilutions: the
rabbit pAb specific for G
o and G
i3 (1:2000) was from Euromedex
(Souffelweyersheim, France); the rabbit pAb AS for G
i1 and -2 (1:200) was a generous gift from Professor P. Mazancourt (Hôpital
Raymond Poincaré, Garches, France) the rabbit pAb raised
against the C-terminal part of G
i3 (1:2500) was from Calbiochem
(Meudon, France); the rabbit pAbs against G
t transducin K-20
(1:500), G
1 C-16 (1:100), and G
2 A-16 (1:500), and the mAb raised
against the c-myc epitope (1:1000) were purchased from Santa Cruz
Biochemicals (Santa Cruz, CA). Membranes were then washed in
Tris-buffered saline containing 0.1% Tween 20 and probed for 90 min
with either a donkey anti-rabbit IgGs pAb (1/2000, Amersham), or a goat
anti-mouse IgGs pAb (1/2000, Santa Cruz Biochemicals), and then
revealed by enhanced chemiluminescence Western detection (ECL; Amersham).
For immunofluorescence labeling of the c-myc epitope in MDCKts.src
cells stably transfected by the CD8-ct-
ARK fusion gene, kidney
cells were grown on autoclaved glass slides (VWR, West Chester, PA) and
processed at 70 to 80% confluence. Cells were rinsed with
phosphate-buffered saline (PBS) at room temperature, fixed for 30 min
in phosphate buffer containing 4% formaldehyde, and washed 3 times in
phosphate buffer. Cells were then made permeable with Triton X-100 in
PBS for 10 min, blocked with 3% goat serum at room temperature for 30 min, and incubated overnight at 4°C with primary mAb anti-c-myc
(Santa Cruz, 1:500). Then, cells were rinsed with PBS and incubated for
1 h at room temperature with a fluorescein-conjugated goat
anti-mouse IgG (1:150) from Immunotech (Roissy, France). Stained
cells were rinsed again in PBS, fixed with a solution of glycerol in
PBS (9:2, v/v), and viewed by conventional epifluorescence on an
Olympus BH2 microscope (Olympus, Tokyo, Japan).
Total RNA was extracted from parental and G
t-transfected MDCKts.src
cells (clones 3-5) using the Trizol reagent (Invitrogen). RT-PCR was
performed using the Superscript detection kit (Invitrogen) according to
the manufacturer's protocol. The experiment comprised one cycle of 30 min at 55°C and one of 2 min at 94°C, followed by 35 cycles of
30 s at 94°C, 30 s at 65°C, and 1 min at 72°C, with a
final extension time of 10 min at 72°C. For human G
t, we used the
following sense and antisense oligonucleotides:
5'-GGCAGACACTATCGAGGAGGGCACGATGCC-3' and
5'-GAGGTGC-GCCTTCTTGACCTTCTCGAAGAAGAC-3'. PCR products were then
separated by electrophoresis in 1.5% agarose gel and detected under UV
light. The G
t primers (Invitrogen) amplified the 519-base-pair fragment specific of the human G-protein transcript.
Glyceraldehyde-3-phosphate dehydrogenase messengers (574 base pairs)
were also analyzed by RT-PCR to verify the integrity of RNA preparations.
Collagen Invasion and Cell Adhesion 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 and allowed to gel. Cells were harvested using
Moscona buffer and trypsin/EDTA and were seeded on top of collagen
gels. Cultures were incubated for 24 h at the indicated
temperature in the presence or absence of the indicated effectors. The
depth of cell migration inside the gels was measured with the use of an
inverted microscope (Vleminckx et al., 1991
). Invasive and superficial
cells were counted in 12 fields of 0.157 mm2. The
invasion index is the number of cells invading the gel expressed as a
percentage of the total number of cells.
Peptides and Reagents.
HGF scatter factor was a generous
gift from Professor Paolo Comoglio (University of Turin, Italy).
Purified recombinant human ITF and hpS2 were a generous gift from Dr.
L. Thim (Novo Nordisk, Bagsvaerd, Denmark) and Professor B. Westley
(University of Newcastle upon Tyre, UK). Leptin was from R & D
Systems Europe (Oxford, United Kingdom). PTx, PAF, thrombin, and
phenylmethylsulfonyl fluoride were from Sigma (Saint Quentin Fallavier,
France). The PAR-1 agonist TRAP (SFLLRN) was from Bachem Biochimie
(Voisins-le-Bretonneux, France). Collagen type I was from Upstate
Biotechnology (Lake Placid, NY).
 |
Results and Discussion |
Negative Control of Cellular Invasion Pathways by PAF-R, PAR-1, and
the Activated Forms AG
o/i of the PTx-Sensitive G-Proteins.
We
have recently shown that PAF receptors exert a negative control on src-
and HGF-induced cellular invasion in kidney and colonic epithelial
cells (Kotelevets et al., 1998
). We first re-examined the effect of PTx
on positive and negative invasion pathways induced by G-protein
-coupled receptors and other transduction elements, including Met, src,
and the activated form of phosphatidylinositol 3-kinase (PI3'-K
)
(p110*). As shown in Fig. 1A, 200 ng/ml
PTx fully reversed the inhibitory effect of PAF on invasion induced by
src in MDCKts.src cells incubated at the permissive temperature 35°C
(invasion index, 7.5 ± 0.6%). The A protomer of PTx is a single
peptide that ADP-ribosylates cysteine residues of membrane G
o/i
subunits, blocking the heterotrimeric complex in the inactive GDP-bound
state and preventing dissociation of G
and G
. The toxin
uncouples the activation of the heterotrimeric complex by PAF-R. In
contrast, the same concentrations of PTx (20-200 ng/ml) produced the
inverse effect and abolished HGF-induced cellular invasion in
MDCKts.src cells incubated at the nonpermissive temperature 40°C
(Fig. 1B), directly implicating substrate G-protein
-subunits and
released G
subunits in signaling through the Met tyrosine kinase.
Similarly, other signals mediated by tyrosine kinase-associated receptors, including epidermal growth factor, insulin-like growth factor-1, basic fibroblast growth factor, erythropoietin Epo
receptors, and insulin receptors, can be inhibited or induced by PTx
treatment and G
sequestrant because of direct and indirect
associations with G
i, G
, and G
subunits (Luttrell et al.,
1997
; Hallak et al., 2000
). Indirect cross-talk and
trans-activation of the epidermal growth factor receptors
was also induced by the G-protein-coupled receptors for thrombin and
lysophosphatidic acid (Prenzel et al., 1999
). Note that PTx did not
abolish invasion induced by the tyrosine kinase src and constitutively
activated PI3'-K
in MDCKts.src and MDCKp110* cells (Fig. 1, C and
D), leptin, and TFFs in kidney and colonic epithelial cells. These
findings raise the possibility that the PTx-sensitive trimeric subunits
on one hand and the G
subunits on the other are selectively and,
respectively involved in negative and positive invasion pathways. To
explore this possibility, we stably transfected MDCKts.src cells with
GTPase-deficient mutants of G
o/i encoding-activated forms (AG
o/i)
of these four G
PTx-sensitive subunits. To confirm that AG
o/i
forms are expressed in MDCKts.src cells, immunoblot analysis in Fig.
2 identified several positive clones
overexpressing G
o (clones 1 and 4), G
i1 (clone 6), G
i2 (clones
1, 7, 8, and 10) and G
i3 (clones 1-3). For example, densitometry analysis revealed that the G
i3 protein was overexpressed 2- and 3-fold in clones 1 and 2, respectively. As shown in Fig.
3, A and B, stable expression of AG
o/i
completely abolished cellular invasion induced by HGF and src,
suggesting that all four PTx-sensitive G
subunits are
candidate-signaling elements in the negative control of cellular
invasion by activated PAF-R. To gain insight into this possibility, we
next determined whether PAF and PAR-1 thrombin receptors have similar
activities on cellular invasion induced by the trefoil factors pS2 and
ITF. The PAR-1 receptor is a seven-transmembrane domain
G-protein-coupled receptor that is also connected with the
PTx-sensitive proteins G
o/i (Seiler et al., 1996
). As shown in Fig.
4, cellular invasion induced by ITF in
kidney MDCKts.src and colonic PCmsrc epithelial cells is resistant to
PAF (A and B), and the same situation is observed in kidney
MDCKts.src-pS2 and colonic HCT8/S11-pS2 epithelial cells stably
transfected by the trefoil factor pS2 (C and D). In contrast, the PAR-1
agonist TRAP abolished ITF- and pS2-induced invasion in the four
models, suggesting that the G
o/i subunits exert a differential
control on invasion. In agreement with this interpretation, we observed that thrombin or the exogenous agonist TRAP completely inhibited cellular invasion induced by HGF in MDCKts.src cells and activated PI3-K in MDCKp110* cells in a PTx-sensitive manner (data not shown). Pertussis toxin (200 ng/ml) also completely reversed the inhibitory effect of thrombin or TRAP on HGF-, ITF-, and src-mediated invasion in
MDCKts.src cells (not shown). We therefore examined the relative contribution of the activated forms of G-proteins AG
o/i on the negative control exerted by PAF-R and PAR-1 receptors on ITF- and
pS2-induced invasion. As shown in Fig. 5,
all forms of AG
o/i blocked invasion induced by leptin and HGF in
MDCKts.src cells. In contrast, only MDCKts.src cells stably transfected
by AG
i3 (clones 1 and 2) seem incapable of responding to ITF and
pS2, suggesting that only this G
i3 subunit can mediate the negative control exerted by activated PAR-1 receptors on cellular invasion induced by the TFFs.

View larger version (27K):
[in this window]
[in a new window]
|
Fig. 1.
Differential control by PTx of cellular invasion
pathways induced by src, HGF/Met, and the activated form p110* of
PI3'-K in MDCK cells. MDCKts.src cells were incubated for 24 h
in the presence or absence of the indicated effectors [PAF (0.1 µM)
or HGF (10 units/ml)] alone or with increasing concentration of PTx
(20-200 ng/ml). The percentage of invasive cells in collagen type I
gels was measured either at the permissive temperature 35°C for src
activation (A and C) or at the restrictive temperature 40°C (B). D,
PTx was tested at 37°C in kidney MDCK cells stably transfected by the
constitutively activated form of PI3'-K (MDCKp110*). Data are
means ± S.E. of three or four separate experiments.
|
|

View larger version (68K):
[in this window]
[in a new window]
|
Fig. 2.
Expression by Western blot of the constitutively
activated forms of the AG o/i1 to -3 subunits in stably transfected
MDCKts.src cells. Parental MDCKts.src cells (P) were stably transfected
by the pcDNA3.1 expression vector recombined with the GTPase-deficient
forms of G o-Q205L (AGo), G i1-Q204L, G i2-Q204L, and
G i3-Q204L (AGi1-3). Positive MDCKts.src clones expressing given
AG subunits were identified by immunoblot analysis using the
antibodies specified under Materials and Methods.
Immunoblots were revealed by the ECL Western detection system and
quantified by the use of a densitometer. The G subunits migrate at
41 kDa (arrow). Results are representative of one other independent
experiment.
|
|

View larger version (26K):
[in this window]
[in a new window]
|
Fig. 3.
Negative control of src- and HGF-induced invasion of
collagen gels by constitutively activated forms of the AG o/i1-3
subunits in MDCKts.src cells. Cellular invasion was compared in
MDCKts.src cells before (parental cell line) and after stable
expression of the constitutively activated forms of G o (clone 1),
G i1 (clone 6), G i2 (clone 1), and G i3 (clone 2). The effect of
the AG o/i1-3 subunits was tested on invasion induced by HGF at
40°C in MDCKts.src cells (A) and by src in MDCKts.src cells incubated
at the permissive temperature 35°C (B). The percentage of invasive
cells was determined as described under Materials and
Methods. Data are means ± S.E. of three separate
experiments.
|
|

View larger version (35K):
[in this window]
[in a new window]
|
Fig. 4.
Differential signaling of PAF-R and PAR-1 thrombin
receptors toward cellular invasion pathways induced by trefoil peptides
in kidney and colonic epithelial cells. The effect of PAF (0.1 µM)
and the PAR-1 agonist TRAP (10 µM) was tested on invasion of collagen
gels induced by ITF (100 nM) in kidney MDCKts.src and colonic PCmsrc
epithelial cells (A and B). Results were compared with the effects of
PAF and TRAP on constitutive cellular invasion induced by
overexpression of the pS2 cDNA in kidney MDCKts.src-pS2 and human
colonic HCT/8S11-pS2 epithelial cells (C and D). The percentage of
invasive cells was determined as indicated under Materials and
Methods. Data are means ± S.E. of three separate
experiments.
|
|

View larger version (32K):
[in this window]
[in a new window]
|
Fig. 5.
Differential signaling of activated AG o/i1-3
subunits toward cellular invasion pathways induced by leptin, HGF, and
trefoil factors in kidney and colonic epithelial cells. Cellular
invasion induced by leptin (100 ng/ml), HGF (10 units/ml), and the
trefoil factors ITF or pS2 (100 nM) was measured at 40°C in
MDCKts.src cells before (parental cell line) and after stable
expression of the constitutively activated forms of G o (clones 1 and
4), G i1 (clone 6), G i2 (clone 1), G i3 (clones 1 and 2). The
percentage of invasive cells was determined as indicated under
Materials and Methods. Data are means ± S.E. of
three or four separate experiments.
|
|
Depletion of Free G
Subunits by the 
-Sequestering
Peptides ct-
ARK and G
t: Consequences for Cellular Invasion and
Adhesion.
It is becoming increasingly evident that signaling
through the G
heterodimers and the individual
and
subunits released upon activation of PTx-sensitive G-proteins G
o/i
plays a key role in multiple-signaling pathways. For example,
inhibition or activation of specific isoforms of adenylate cyclase
(Hamm, 1998
), G-protein-coupled receptor kinases GRK1 to -3 involved in
receptor desensitization (Pitcher et al., 1992
; Carman et al., 2000
),
membrane-targeting of G
subunits and binding to pleckstrin homology
domains (Touhara et al., 1994
), G
exchange between G
i-coupled
and G
q-coupled receptors (Quitterer and Lohse, 1999
), src- and
ras-dependent activation of mitogen-activated protein kinases,
G
-sensitive PI3'-K
and Jun kinases, and phospholipase C-
are concerned with G
signaling (Luttrell et al., 1996
;
Lopez-Ilasaca et al., 1998
; Rickert et al., 2000
). Recently, release of
G
from activated insulin-like growth factor I receptor
(Hallak et al., 2000
) and direct interactions between G
and the
Rho family of GTPases, Rho and Rac, have been documented (Harhammer et
al., 1996
; Ueda et al., 2000
). The role of G
signaling in either
cellular invasion or adhesion is unknown. Accordingly, we have
established MDCKts.src cell lines stably expressing the G
sequestrant vector comprising the CD8 receptor and ct-
ARK.
As shown in Fig. 6, immunoblot analysis
identified the ectopic expression of the myc epitope-tagged
CD8-ct-
ARK protein in MDCKts.src-ct-
ARK cells (clones 3-5).
The signal was not detected by immunoblotting in parental MDCKts.src
cells and their transfected counterparts, the MDCKts.src-ct-
ARK
cells (clones 1 and 2). Strong expression of the G
subunit
inhibitor was confirmed by indirect immunofluorescence in
MDCKts.src-ct-
ARK clone 3, whereas a much weaker signal was present
in MDCKts.src-ct-
ARK clone 1 cells, at barely detectable levels. We
therefore examined the invasive properties of the low- and
high-expressing MDCKts.src-ct-
ARK cells (clones 1 and 3). In the
MDCKts.src-ct-
ARK clone 1 expressing very low levels of
CD8-ct-
ARK, cellular invasion induced by HGF was reduced by 50%
(Fig. 7A) compared with nontransfected
parental MDCKts.src cells. This residual activity was still sensitive
to PAF inhibition, whereas src-induced cellular invasion was unaffected in MDCKts.src-ct-
ARK clone 1 (Fig. 7B). In contrast, high
expression of the G
scavenger leads to total inhibition of the
invasive phenotype controlled by HGF and/or src in the
MDCKts.src-ct-
ARK clone 3 (Fig. 7, A and B). These findings suggest
that the G
subunits are required as potential links between
G-protein-coupled receptors and positive cellular invasion pathways
induced by the src and Met oncogenes. Similarly, the G
inhibitor
CD8-ct-
ARK abolished leptin-, pS2- and ITF-induced cellular
invasion in MDCKts.src cells (not shown). Over the last few years, a
number of interesting connections between invasion pathways and
cellular adhesion to extracellular matrix proteins have emerged.
Increasing evidence shows multiple collaborations between cell-cell and
cell-substratum interactions, assembly of the actin cytoskeleton, and
signal transduction pathways involved in invasion and metastasis. These
interactions, together with recent reports suggesting activation by
G
subunits of the Rho family GTPase including Ras, Rho, and Rac
(Harhammer et al., 1996
; Ueda et al., 2000
), prompted us to examine the
adhesive properties of MDCKts.src cells transfected with the G
sequestrant CD8-ct-
ARK. The interdependence of these pathways is
further illustrated by the sequential role of Rac and Rho in the
initiation of new adhesion sites and their maturation into focal
adhesions.

View larger version (76K):
[in this window]
[in a new window]
|
Fig. 6.
Western blot and immunofluorescence analysis of the
G sequestrant peptide ct- ARK in MDCKts.src cells stably
transfected the CD8-ct- ARK expression vector. Top, expression of
the myc-tagged -adrenergic receptor kinase (ct- ARK) peptide in
MDCKts.src cells transfected by the CD8-ct- ARK vector (clones
1-5). Immunoblotting of total cell lysates was performed using the
c-myc mAb and revealed by the ECL Western detection system. Bottom,
stable expression of the G sequestrant was analyzed by indirect
immunofluorescence in MDCKts.src-ct- ARK cells (clones 1 and 3).
Cells were processed with the mAb against the c-myc epitope, followed
by fluorescein-conjugated goat anti-mouse IgG as secondary antibody, as
described under Materials and Methods.
|
|

View larger version (34K):
[in this window]
[in a new window]
|
Fig. 7.
Negative control of src- and HGF-induced invasion of
collagen gels by the G sequestrant peptide ct- ARK in
MDCKts.src cells stably transfected with the CD8-ct- ARK expression
vector. Cellular invasion index was determined in parental and
ct- ARK-transfected MDCKts.src cells (clones 1 and 3) incubated at
the nonpermissive temperature 40°C for src activation (A) or at the
permissive temperature 35°C (B). MDCKts.src-ct- ARK cells were
treated with 10 units/ml HGF alone or with 0.1 µM PAF and compared
with their respective control cells (no effector). The percentage of
invasive cells was determined as described under Materials and
Methods. Data are means ± S.E. of three separate
experiments.
|
|
As shown in Fig. 8, A and B,
sequestration of the G-protein 
subunits impaired cell adhesion
and spreading to collagen gels and filopodia formation in
MDCKts.src-ct-
ARK clone 3 activated by the HGF/Met and src
oncogenes compared with the low-expressing ct-
ARK clone 1. Cell
adhesion to extracellular matrix proteins is responsible for cell
spreading through integrin receptors, which promotes integrin
clustering and cytoskeletal reorganization and induces cells to spread.
Disruption of the gene encoding the G-protein
-subunit impaired the
regulation of the actin cytoskeleton at cell-yeast particle adhesion
sites during phagocytosis and abrogated chemotaxis in the
Dictyostelium discoideum amebae (Peracino et al., 1998
),
suggesting that G
is intimately involved in signal transduction
networks linking cytoskeletal responses to chemoattractants. The
G-proteins 
may also be associated with adhesion and/or morphological changes (Hansen et al., 1994
). In the present study, inhibition of G
signaling effectively induced cell rounding and
alterations of spreading in MDCKts.src-ct-
ARK clone 3 that was
refractory to activation of src and Met (Fig. 8B).

View larger version (94K):
[in this window]
[in a new window]
|
Fig. 8.
Effect of overexpressed G sequestrant peptide
CD8-ct- ARK on adhesive properties and spreading of MDCKts.src cells
on collagen gels. Adhesion to collagen type I gels was compared in
MDCKts.src-ct- ARK cells (clones 1 and 3 in A and B, respectively)
incubated for 24 h at 35°C or 40°C in the presence or absence
of HGF (10 units/ml). Cell were seeded at the density of
105 cells/ml. The white arrows in A underline the
src-dependent filopodia formation in MDCKts.src-ct- ARK clone 1 incubated in the presence or absence of HGF. In contrast, the
MDCKts.src-ct- ARK clone 3 showed a remarkable reduction of
filopodia protrusions and cellular spreading induced by src (B). Most
interestingly, ct- ARK caused shape change and cellular rounding in
MDCKts.src-ct- ARK clone 3 incubated at the nonpermissive
temperature for src activation (B), leading to a loss of cell adhesion,
even in the presence of the Met activator HGF (black arrows).
|
|
As shown in Fig. 9, overexpression of the
G
sequestrant vector G
t transducin also prevented cellular
invasion induced by HGF, leptin, and pS2, either alone (A) or combined
with activation of the src oncogene (B) in MDCKts.src-G
t cells
(clone 5). Similar results were obtained in MDCKts.src-G
t cells
(clones 3 and 4) established from transfected MDCKts.src cells
expressing the human G
t transducin protein by immunoblot (data not
shown) and RT-PCR (Fig. 9, inset). Because G
subunits have been
implicated in the regulation of Rho GTPases that are dynamic regulators
of the actin cytoskeleton (Harhammer et al., 1996
; Hall, 1998
; Ueda et al., 2000
), including the Rho/Rac interplay, our data on the 
sequestering peptides ct-
ARK and G
t transducin are consistent with the direct implication of G
subunits in several signaling pathways involved in cellular invasion and switch between cellular rounding/retraction and spreading/cellular adhesion.

View larger version (28K):
[in this window]
[in a new window]
|
Fig. 9.
Negative control of src- and HGF-induced invasion of
collagen gels by the G sequestrant peptide G t in MDCKts.src
cells stably transfected by the G t transducin expression vector.
Cellular invasion index was determined in parental and
G t-transfected MDCKts.src cells cultured at 40°C (A) or at the
permissive temperature 35°C for src activation (B). Stably
transfected MDCKts.src-G t cells (clone 5) were treated with either
10 units/ml HGF, 100 ng/ml leptin, or 100 nM pS2 and compared with
their respective control cells (no effector) and parental MDCKts.src
cells incubated in the presence or absence (Control) of 100 ng/ml
leptin. The percentage of invasive cells was determined as described
under Materials and Methods. Data are means ± S.E.
of three separate experiments performed in clone 5 and are
representative of another experiment performed in clones 3 and 4.
|
|
Activation of Cellular Invasion by Coexpression of G
Subunits.
Our data on the G
sequestering proteins ct-
ARK
and G
t support the hypothesis that free G
endogenous dimers
released from activated PTx-sensitive G-proteins are required for
initiation and activation of positive cellular invasion pathways
controlled by oncogenes and epigenetic factors. To explore further such
a possibility, we next examined the effect of ectopic overexpression of
G
1 alone or combined with G
2 in MDCKts.src cells. First, we
stably transfected MDCKts.src cells with G
1 and selected four G418-resistant colonies (Fig. 10A).
Immunoblot analysis identified two MDCKts.src-G
1 cells (clones 1 and
3) overexpressing the G
1 protein (36 kDa). As shown in Fig. 10B,
HGF-induced invasion of collagen gels was completely blocked by
overexpression of G
1 in clone 3, whereas src-induced activation was
unaffected (Fig. 10C). Thus, we examined the G
1-dependence of the
negative control exerted by PAF-R and PAR-1 on src-induced invasion. As
a consequence of G
1 overexpression, there was a substantial
attenuation of PAF-R-mediated inhibition of cellular invasion (Fig.
10C), but no significant effect on the negative control exerted by
PAR-1 was found. Our data are therefore consistent with the possibility that overexpression of G
1 alone exerts a selective inhibition of the
PAF-R and HGF/Met signaling pathways. This possibility is consistent
with the recent finding that molecular complexes between G
subunits
and several regulators of G-protein signaling (RGS) have been
identified, because structural domains in RGS proteins exhibit striking
homologies to G
subunits (Snow et al., 1998
; Levay et al., 1999
;
Sowa et al., 2000
). Most importantly, some G
/RGS heterodimers behave
as GTPase-activating proteins for certain G
subunits, such as G
o
(Snow et al., 1998
). The RGS activity and selectivity for PTx-sensitive
or -insensitive G
subunits can be directly or indirectly determined
through their interactions with different G
subunits,
G-protein-coupled receptors, or other binding partners. For example,
RGS1, RGS3, RGS4, and GAIP stimulate the GTPase activity of G
i
family members but are ineffective against G
s. Thus, overexpression
of G
1 can selectively mimic, at least in part, PTx-induced
deactivation of the PAF-R and Met signaling pathways involved in
cellular invasion, as shown in Fig. 1.

View larger version (28K):
[in this window]
[in a new window]
|
Fig. 10.
Effect of overexpressed G 1 subunit on positive
and negative invasion pathways in stably transfected MDCKts.src-G 1
cells. A, expression of the G 1 subunit was analyzed by Western blot
in parental MDCKts.src cells (P) and transfected MDCKts.src-G 1 cells
(G418-resistant clones 1-4), using the G 1 pAb. The G 1 subunit
migrates at 36 kDa (arrow). Cellular invasion was then quantified in
the MDCKts.src-G 1-positive clone 3 incubated at 40°C (B) or
35°C (C). Parental MDCKts.src and MDCKts.src-G 1 cells were treated
with either HGF (10 units/ml), PAF (0.1 µm), or TRAP (10 µm) and
compared with their respective control cells (no effector). The
percentage of invasive cells was determined, as described under
Materials and Methods. Data are means ± S.E. of
three separate experiments.
|
|
Individual G
or G
subunits exert divergent signaling functions,
depending on the functional status and relative distribution of the
resident G
and G
subunits and dimers interacting with a given
signal transduction system. In the G
complexes, the G
subunits
consist of seven subtypes with high amino acid sequence identity from
80 to 90% for G
1- to -4 and 52% identical for G
5 (Watson et
al., 1994
). Thus, G
2 subunit was expressed together with G
1 in
MDCKts.src cells after cotransfection of MDCKts.src-G
1 cells (clone
3) with the G
2 (3 µg) and pcDNA3.1/Hygro (0.3 µg) expression
vectors and selection for 2 weeks in the presence of hygromycin (0.5 mg/ml). As shown in Fig. 11, immunoblot
analysis identified two MDCKts.src-G
2-positive cells (clones 4 and
11) expressing the ectopic G
2 protein (7 kDa) after subsequent
transformation of MDCKts.src-G
1 cells by the G
2 vector. The G
2
subunit protein was not detected in MDCKts.src cells. Coexpression of
both G
1 and G
2 subunits in MDCKts.src cells resulted in a
remarkable induction of cellular invasion in MDCKts.src-G
1
2 cells
(Fig. 11B; clones 4 and 11) and potentiation of invasiveness induced by
HGF (Fig. 11, B and C), and src (Fig. 11C). Furthermore, invasion induced by overexpressed G
1
2 proteins in MDCKts.src cells
(invasion index, 5.2% and 4.8% in clones 4 and 11, respectively) is
abrogated by activation of the PAF-R and PAR-1. Our data therefore
indicate that formation of the G
1
2 complex in MDCKts.src-G
cells is associated with induction and potentiation of cellular
invasion pathways. This is in agreement with our demonstration that
cellular invasion induced by src, Met, and epigenetic factors was
abrogated by sequestration of free G
subunits by ct-
ARK and
G
t. Thus, PTx neutralizes HGF-induced invasion (Fig. 1) and reverses
inhibition of cellular invasion controlled by activated PAF-R and
PAR-1, suggesting that PTx abolishes 1) G
-signaling pathways
involved in the induction of cellular invasion by HGF, and 2)
G
o/i-mediated inhibition of cellular invasion induced by PAF/TRAP.
Alternatively, it should be stressed that G
i subunits are also
activators of invasion pathways controlled by src and PI3'-K
,
for example.

View larger version (38K):
[in this window]
[in a new window]
|
Fig. 11.
Positive control of cellular invasion after ectopic
coexpression of G 1 and G 2 subunits in stably transfected
MDCKts.src-G 1 2 cells. Reversion by PAF-R and PAR-1 activation. A,
expression of the G 2 subunit in MDCKts.src-G 1 cells (P)
subsequently transfected by the G 2 vector. Immunoblotting of total
cell lysates was performed using the affinity-purified rabbit pAb G 2
(A-16) and revealed two positive MDCKts.src-G 1 2 cells (clones 4 and 11) by the ECL Western detection system. B and C, cellular invasion
index was determined in parental MDCKts.src cells incubated in the
presence or absence of 100 ng/ml leptin and in MDCKts.src-G 1 2
cells (clones 4 and 11) incubated at 40°C (B) or at the permissive
temperature 35°C for src activation (C). Stably transfected
MDCKts.src-G 1 2 cells and parental MDCKts.src cells were treated
with either 10 units/ml HGF, PAF (0.1 µM), or the PAR-1 agonist TRAP
(10 µM) and compared with their respective control cells (no
effector). The percentage of invasive cells was determined as described
under Materials and Methods. Data are means ± S.E.
of three separate experiments.
|
|
In conclusion, our data indicate that the redistribution of
heterotrimeric G-proteins from activated PAF and PAR-1 receptors lead
to a negative control of cellular invasion via the activation of
G
o/i subunits in response to multiple invasion pathways induced by
oncogenes and epigenetic factors (Fig.
12). Thus, the G
i/o subunits exert a
dominant invasion-suppressor role in our assay. In contrast, G
subunits that are liberated from PTx-sensitive G
subunits are
critical mediators and links between positive cellular invasion and
adhesion pathways. In this scenario, G
dimers can induce
activation of multiple signaling pathways that are critical components
of cellular invasion, namely Rho-like small G-proteins, the
/
isotypes of PI3'-K, and the phospholipase C/protein kinase C cascade
(Rickert et al., 2000
). The individual functions of the paired G
proteins and the PTx-insensitive G
subunits associated with PAF-R
and PAR-1 may provide an alternative pathway by which these
heterotrimeric G-proteins may exert opposing effects on cellular
invasion (i.e., positive and negative signals). Precisely how the
balance between these positive and negative signals is regulated in the
integration of the cellular responses remains to be elucidated.
Disruption of this equilibrium might have important biological
consequences, because we have shown that PAF and PAR-1 receptors exert
a dominant negative function on cellular invasion. Deactivation of
G
o/i subunits by PTx in the present study can be physiologically
induced and mimicked by the RGS proteins that negatively control these
PTx-sensitive G-proteins and G
z, G
12/13 subunits, and PAF-R
phosphorylation as well. In contrast, we found that overexpressing
G
subunits was sufficient to induce invasiveness. Thus, RGS
signals might abolish the dominant negative control exerted by G
o/i
subunits on cellular invasion and trigger a permissive action on
positive invasion pathways governed by G
. These results are
analogous to the negative cross-talk exerted by G
o on the G
q
pathway and its downstream effectors (Hajdu-Cronin et al., 1999
).
Abnormalities in the expression, protein structure, and constitutive
activation of G-protein subunits (G
, G
i2, G
i3) have been
reported in human pathologies, including tumors of the ovary,
neuroendocrine tumors, and experimental models of colonic neoplasms
induced by azoxymethane in rats (Lyons et al., 1990
; Bolt et al., 1998
;
Farfel et al., 1999
). Recently, mutants of the exchange factor cdc24
for the Rho-like GTPase Cdc42 defective in binding to the G-protein
G
subunit have been described (Nern and Arkowitz, 1998
) in the
context of the cytoskeletal reorientation. Therefore, activation or
inactivation of the molecular components of the PTx-sensitive and
-insensitive trimeric G-proteins in familial and sporadic tumors
(Rodrigues et al., 2001
) should be considered in view of their
differential localization and function in basolateral and cytoskeletal
domains, including focal adhesion and stress fibers in polarized
epithelial cells in kidney and intestine (Hansen et al., 1994
; Bolt et
al., 1998
; Saha et al., 1998
). Further studies will shed more light for
understanding the relative contribution and functioning of heterotrimeric G-proteins G
and G
in positive and negative signaling pathways involved in cellular adhesion and invasion in the
context of neoplasia and metastasis.

View larger version (34K):
[in this window]
[in a new window]
|
Fig. 12.
Model depicting the negative and positive invasion
pathways induced by activated PAF-R and PAR-1 in kidney and colonic
epithelial cells. Activation of G o/i subunits by serpentine PAF-R
and PAR-1 exerts a dominant negative control ( ) on cellular invasion
pathways governed by src, HGF/Met, the epigenetic factors leptin and
TFFs, and the G subunits (+) that are released from
these PTx-sensitive G-proteins. The G subunits, on the other
hand, are invasion promoters, according to their connections with src,
phospholipase C- , PI3'-kinases, and the Rho-like GTPases that induce
the formation of filopodia and lamellipodia (Cdc42 and Rac), focal
adhesions, and stress fibers (Rho). The G o/G i downstream targets
remains to be identified (?) in the context of the integration of the
positive and negative signaling pathways implicated in cellular
spreading, migration, local invasion, and distant metastasis.
|
|
We are grateful to Professor P. Comoglio, Professor P. de
Mazancourt, Dr. E. Peralta, Dr. L. Thim, Professor R. Weitzker, and
Professor B. Westley for providing reagents.
This work was supported by the Institut National de la
Santé et de la Recherche Médicale, Research Grants from la
Fondation pour la Recherche Médicale, France (to S.F.),
l'Association de la Recherche sur le Cancer, France (to C.G. and
S.E.), the Fortis Bank, Verzekeringen and the FWO-Vlaaderen, Belgium.
S.F. and K.R. contributed equally to this work.
Dr. Christian Gespach, INSERM
U482: Signal Transduction and Cellular Functions in Diabetes and
Digestive Cancers, Hôpital Saint-Antoine, 184 rue du Faubourg
Saint-Antoine, 75571 Paris Cedex 12, France. E-mail:
gespach{at}st-antoine.inserm.fr
PAF, platelet-activating factor;
TFF, trefoil
factor;
PAF-R, platelet-activating factor receptor;
MDCK, Madin-Darby
canine kidney;
PTx, pertussis toxin;
ITF, intestinal trefoil factor;
AGo/i, activated forms of G
0, G
i1, G
i2, and G
i3;
ct-
ARK, C-terminal end of the
-adrenergic receptor kinase;
G
t, transducin
G
subunit;
PAR-1, protease-activated receptor-1;
TRAP, thrombin
receptor-activating peptide;
HGF, hepatocyte growth factor;
RT, reverse
transcriptase;
PCR, polymerase chain reaction;
pAb, polyclonal
antibody;
mAb, monoclonal antibody;
PBS, phosphate-buffered saline;
ECL, enhanced chemiluminescence;
PI3'-K
, activated form of
phosphatidylinositol 3-kinase;
RGS, regulators of G-protein
signaling.