(Received for publication, April 1, 1997, and in revised form, June 3, 1997)
From the Dipartimento di Genetica, Biologia e Chimica
Medica, Universita' di Torino, 10126 Torino, Italy, the
Dipartimento di Psicologia,
Università di Roma "La Sapienza," Italy, the
¶ Laboratoire d'Hématologie INSERM U217, Grenoble, France,
the
INSERM U452, Nice, France, and the ** Laboratorio di
Ultrastrutture, Istituto Superiore di Sanità, Roma, Italy
A panel of antibodies to the IIb
3 integrin
was used to promote adhesion of Chinese hamster ovary cells transfected
with the
IIb
3 fibrinogen receptor. While some
IIb
3
antibodies were not able to induce p125 focal adhesion kinase (p125FAK)
tyrosine phosphorylation, all the antibodies equally support cell
adhesion but not spreading and assembly of actin stress fibers. Absence of stress fibers was also obtained by plating on antibodies directed to
the hamster
1 integrin. In contrast, cells plated on matrix proteins
spread organizing actin stress fibers. Treatment with phorbol esters
phorbol 12-myristate 13-acetate (PMA) induced cells to spread on
antibodies-coated dishes but not to organize actin in stress fibers.
The combination of PMA and cytotoxic
necrotizing factor 1 (CNF1), a specific Rho
activator, induced cell spreading and organization of stress fibers.
PMA or the combination of PMA and CNF1 also increases tyrosine
phosphorylation of p125FAK in response to antibodies that were
otherwise unable to trigger this response. These data show that: 1)
matrix proteins and antibodies differ in their ability to induce
integrin-dependent actin cytoskeleton organization (while
matrix induced stress fibers formation, antibodies did not); 2) p125FAK
tyrosine phosphorylation is insufficient per se to trigger
actin stress fibers formation since antibodies that activate p125FAK
tyrosine phosphorylation did not lead to actin stress fibers assembly;
and 3) the inability of anti-integrin antibodies to trigger stress
fibers organization is overcome by concomitant activation of the
protein kinase C (PKC) and Rho pathways; PKC activation leads to cell
spreading and Rho activation is required to organize actin stress
fibers.
Integrins are transmembrane adhesive receptors formed by and
subunits, which connect the matrix to the cytoskeletal structure of
the cell (1). Cell matrix adhesion triggers the reorganization of cell
shape and also determines important cellular functions as cell cycle
entry and differentiation (reviewed in Refs. 2 and 3). Integrins
trigger cytoplasmic signals such as pH variations, Ca2+
influx, potassium channels activation (4-6), and tyrosine
phosphorylation of cytoplasmic proteins (for review, see Ref. 7).
Tyrosine phosphorylation of the p125 focal adhesion kinase
(p125FAK),1 which is
specifically localized in the focal contacts (8, 9), is an early event
in integrin signaling in cells plated on matrix proteins (10-13) and
by clustering of
1 integrins on cells adherent to plastic dishes
(14) or on cells in suspension (12).
Integrin-mediated adhesion induces cytoskeletal organization, leading to actin stress fibers formation (15). The relevance of tyrosine phosphorylation to the assembly of focal adhesion and actin stress fibers has been determined by using phosphotyrosine kinase inhibitors. Herbimycin A and genistein, two known inhibitors of kinase activity, prevent focal adhesion formation and actin stress fibers assembly in cells plated on fibronectin (10, 12). Moreover treatment of cells with phosphotyrosine phosphatase inhibitors leads to increased integrin-dependent organization of focal adhesion and actin stress fibers (16). In addition to tyrosine kinases, small GTPases have been involved in the assembly of the actin cytoskeleton in response to neuropeptides and growth factors (17). In the Rho-dependent pathway, the tyrosine kinases inhibitor genistein blocks the assembly of stress fibers, suggesting a role for tyrosine kinases downstream to Rho (18). Several tyrosine kinases located in the focal adhesions, including Src and p125FAK, are potential candidates (7, 15). The exact role of p125FAK tyrosine phosphorylation in the assembly of actin stress fibers induced by the Rho pathway remains to be determined.
To assess the ability of integrins to regulate cell spreading and actin
cytoskeleton organization, Chinese hamster ovary (CHO) cells
transfected with the human IIb
3 receptor were plated on anti-human
IIb
3 or anti-hamster
5
1 antibodies, and the
roles of p125FAK kinase, protein kinase C and small GTPase Rho were analyzed. We report here that antibodies directed either to the
IIb
3 or
5
1 trigger the organization of actin in
filapodia-like structures but do not lead to actin stress fibers
formation. The inability to organize stress fibers on anti-integrin
antibodies is overcome by concomitant activation of PKC and Rho. Since
some of the antibodies induce p125FAK tyrosine phosphorylation while others are ineffective, our data also show that p125FAK tyrosine phosphorylation is not sufficient for stress fibers organization.
Full-length IIb cDNA from
nucleotide 1 to 3198 was ligated into the SmaI pECE
restriction site (19), whereas
3 cDNA from nucleotide 18 to 2585 was introduced between the EcoRI and XbaI pECE
restriction sites (19, 20). The resulting plasmids were electroporated
in CHO cells and clones, selected in the presence of G418 (Life
Technologies, Inc., Gaithersburg, MD), were cultured in Ham's F-12
medium supplemented with 10% fetal calf serum (Hyclone). Before each
experiment, to eliminate the contribution of protein synthesis and
secretion, cells were pretreated 2 h with 20 µM cycloheximide (Sigma), and adhesive assays were all performed in
serum-free medium in the presence of 20 µM cycloheximide
and 1 µM monensin (Sigma) (21).
Table I presents a summary of the monoclonal antibodies (mAbs) used (21-27). All the antibodies were affinity purified on protein A-Sepharose (Pharmacia Biotechnology, Uppsala, Sweden) as described (28), and the purity of the antibodies was higher than 95%.
|
Cells at confluence were detached by gentle treatment with 0.25% trypsin in PBS, washed, and incubated with the indicated concentration of antibodies for 1 h at 4 °C. Cells were then incubated with 10 µg/ml affinity purified fluorescein-labeled secondary antibodies (Sigma) for 1 h at 4 °C and analyzed on the flow cytometer Facs-Star Becton Dickinson, equipped with a 5 watt argon laser at 488 nm.
Immunoprecipitation ofCells at confluence were
metabolically labeled with [35S]methionine and
[35S]cysteine (Tran35S-label, ICN Flow, Costa
Mesa, CA), detergent extracted, and immunoprecipitated as described in
(21). To selectively immunoprecipitate the hamster v subunit, a
polyclonal antibody prepared by injecting rabbits with a synthetic
peptide reproducing the cytoplasmic domain of this subunit was used
(29).
96-well microtiter dishes were coated with different concentrations of fibrinogen (Sigma) at 4 °C by overnight incubation and post-coated for 1 h with 1% bovine serum albumin (BSA, Sigma) in PBS. To evaluate antibody-mediated adhesion, microtiter wells were coated overnight at 4 °C with 10 µg/ml purified rabbit anti-mouse IgG (Sigma), post-coated with BSA for 1 h at 37 °C, and incubated for 2 h at 37 °C with the indicated antibodies. Cells were released from culture dishes by treatment with 5 mM EDTA for 10 min at 37 °C, washed twice with PBS, 1 mM CaCl2, 1 mM MgCl2, and plated in serum-free DMEM, 20 mM Hepes on microtiter wells for the indicated times at 37 °C. Plates were rinsed, and adherent cells were fixed with paraformaldehyde and stained with Coomassie Blue. Cell adhesion was evaluated by reading the absorbance at 540 nm in a microtiter enzyme-linked immunosorbent assay (ELISA) reader (Bio-Rad 450; Bio-Rad).
Analysis of p125FAK Tyrosine Phosphorylation by Antibody-mediated Integrin Clustering or Adhesion to Antibodies-coated DishesCells at confluence were detached in 5 mM EDTA as described above, washed, resuspended in DMEM, 20 mM Hepes containing 10 µg/ml purified primary antibodies for 50 min at 4 °C, washed twice, and then incubated with 25 µg/ml purified rabbit anti-mouse immunoglobulins at 37 °C to induce integrin clustering. The cells were washed and detergent extracted in lysis buffer (1% Nonidet P-40, 150 mM NaCl, 50 mM Tris-Cl, pH 8, 5 mM EDTA, 10 mM NaF, 10 mM Na4P2O7, 0.4 mM Na3VO4, 10 µg/ml leupeptin, 4 µg/ml pepstatin, and 0.1 trypsin inhibitory unit/ml of aprotinin) (all from Sigma) (12).
To analyze adhesion-mediated tyrosine phosphorylation, tissue culture plates of 10 cm diameter were coated either with 10 µg/ml of the indicated antibodies or with 20 µg/ml fibrinogen and postcoated with 1% BSA. As a nonspecific, substrate dishes were coated with 10 µg/ml poly-L-lysine (Sigma). Confluent cells were detached as described above and plated in DMEM, 20 mM Hepes. The cells were incubated for 1 h at 37 °C and then washed and detergent extracted as described above.
Detection of Tyrosine Phosphorylated p125FAK and of p125FAK ProteinSamples containing equal amounts of proteins were immunoprecipitated with polyclonal antibody to p125FAK (FAK-4) (16) and separated by polyacrylamide gel electrophoresis in presence of SDS (SDS-PAGE) in reducing conditions. Proteins were transferred to nitrocellulose, and the blots were incubated overnight in antiphosphotyrosine mAb PY20 (0.3 µg/ml) (Transduction) followed by peroxidase-conjugate anti-mouse IgGs (Sigma) (12, 16). Phosphotyrosil-containing proteins were visualized by the chemiluminescent detection method, ECL (Amersham, UK). Exposure times were set to obtain a linear response. To detect the p125FAK protein, the anti-p125FAK mAb 9/2 (16) was used as primary antibody.
Fluorescence MicroscopyFor immunofluorescence microscopy, acid-washed glass coverslips were coated with 20 µg/ml fibrinogen or 10 µg/ml anti-integrin antibodies overnight. Cells were detached by EDTA treatment, washed, and plated for 3 h in serum-free medium with or without treatments. The cells were then fixed in 3% paraformaldehyde, 60 mM sucrose in PBS for 10 min, and permeabilized for 1 min at 4 °C in Tris-buffered saline containing 0.5% Triton X-100. Actin cytoskeleton was visualized with fluorescein-labeled phalloidin (Sigma). The coverslips were mounted in Mowiol (Aldrich) and viewed on an Olympus BH2-RFCA fluorescence microscope. Micrographs were taken on Kodak 400 film.
PMA, RO 31-8220, CNF1, Calphostin C, and Genistein TreatmentsPhorbol 12-myristate 13-acetate (PMA) (Sigma) was dissolved at 10 mM in DMSO. PKC inhibitor mesylate salt RO 31-8220 (33), a gift of Dr. Hallam (Roche, Hertfordshire, UK), was prepared at 0.5 mM in DMSO and was used at final concentration of 10 µM. Calphostin C (Calbiochem, La Jolla, CA) was solubilized in DMSO at 0.2 mM and used at 100 µM. The treatments with PMA, RO 31-8220, or calphostin C were performed in serum-free DMEM supplemented with 20 mM Hepes, pH 7.4, in the presence of 20 µM cycloheximide and 1 µM monensin for the indicated times. Cytotoxic necrotizing factor 1 (CNF1) was purified as described in (30) and treatments were performed overnight in culture medium. CNF1 was also added during adhesive assays in serum-free DMEM as indicated above. Genistein (Sigma) was dissolved at 25 mg/ml in DMSO and used at the final concentration of 74 µM.
Fluorescence-activated cell sorter
analysis showed that the transfected CHO cells express on their surface
high and comparable levels of both IIb and
3 subunits (Fig.
1A, panels b-d).
Metabolically labeled cell extracts immunoprecipitated with antibodies
to the human
3 subunit or to the
IIb
3 complex indicated that
the human
3 was mostly associate with the
IIb (Fig.
1B, lanes a and c). This was also
confirmed by the fact that the anti-
v antibody, which reacts against
the hamster subunit, immunoprecipitates the
v
1 integrin,
suggesting that the endogenous
v subunit preferentially forms
complexes with this
subunit (Fig. 1B, lane
b). While non-transfected wild-type CHO cells did not adhere to
fibrinogen, transfected cells start to adhere to dishes coated with
0.625 µg/ml fibrinogen, reaching the plateau at 2.5-5 µg/ml (Fig.
1C). A time course experiment showed that maximal adhesion
was obtained 60 min after plating (not shown).
To investigate the ability of the
IIb
3 receptor to trigger p125FAK tyrosine phosphorylation,
transfected CHO cells kept in suspension were incubated with monoclonal
antibodies directed to different epitopes of the
IIb,
3 subunits,
or the
IIb
3 complex, and clustering was induced by the addition
of a secondary antibody (see Table I for
summary of the antibodies) (21-27). Clustering by anti-
3 mAb P37
induced high levels of p125FAK tyrosine phosphorylation, indicating
that integrin aggregation by antibodies was sufficient to trigger this
event. Maximal stimulation of p125FAK tyrosine phosphorylation was
obtained as early as 2 min after clustering, persisted at 10 min, and
decreased at 30 min of incubation (Fig.
2A).
The different antibodies used showed distinct abilities to trigger
p125FAK tyrosine phosphorylation. In addition to mAb P37, anti-IIb
3 mAbs P4.1.1 and P8.2.1 and anti-
3 mAb B212 were all capable of inducing p125FAK tyrosine phosphorylation (see Table I). In
contrast, we found that three antibodies, D33C directed to the
IIb
subunit (Fig. 2B) (22), CS3, and P9.1.1 (not shown) to the
IIb
3 complex, were completely unable to induce p125FAK tyrosine
phosphorylation. The lack of p125FAK tyrosine phosphorylation observed
with these antibodies was reproducibily found in five different
experiments (see also the adhesion experiments presented below).
The ability of different antibodies to trigger p125FAK tyrosine
phosphorylation may be due to the binding properties of the antibodies.
To evaluate if the antibodies used differ in their capacity to induce
IIb
3 clustering in the plasma membrane, immunofluorescence experiments were performed. When cells were stimulated with mAbs P37 or
D33C and secondary antibodies at 4 °C, the
IIb
3 protein appeared uniformly diffused on the cell surface (Fig.
3, panel a). When the
temperature was raised to 37 °C, the antibodies gave rise to
redistribution of the antigen on the cell surface and formation of
patches (Fig. 3, panels e-n), indicating that both mAb P37
and D33C were able to cluster the molecules on the plasma membrane.
Staining with phalloidin showed that actin cytoskeleton reorganizes
under the clustered receptors (not shown). Similar integrin
redistribution and actin organization was found with the different
antibodies used (Table I). Interestingly, cluster formation and actin
reorganization were independent from the ability of the different
antibodies to trigger p125FAK tyrosine phosphorylation since mAbs D33C,
CS3, and P9.1.1 induced actin co-clustering without triggering p125FAK
tyrosine phosphorylation.
p125FAK Tyrosine Phosphorylation Is Induced by Adhesion to Fibrinogen or
To test if lack of
p125FAK tyrosine phosphorylation observed in clustering experiments in
suspended cells was related to the absence of cell spreading, cells
were plated for 1 h on anti-IIb
3 antibodies-coated tissue
culture dishes. The ability of the anti-
IIb
3 antibodies to
trigger p125FAK tyrosine phosphorylation was very similar to that
observed in clustering experiments (Fig.
4; see also Table I). While cells plated
on mAb P37 and mAbs P4.1.1 or P8.2.1 were able to induce a level of
p125FAK tyrosine phosphorylation comparable with that observed on
fibrinogen, the
IIb
3-specific ligand used as positive control
(not shown), mAbs D33C and CS3 were poorly effective. The inability of
the latter two antibodies to trigger p125FAK tyrosine phosphorylation
was also found when plating of cells was protracted for 3 h (not
shown). Adhesion experiments (Fig. 1C) showed that
transfected cells adhere in a similar way either to mAb CS3
(ineffective in the induction of p125FAK tyrosine phosphorylation) or
to mAb P37 (effective in the induction of p125FAK tyrosine
phosphorylation), indicating that the two antibodies were equally
capable of promoting cell attachment. These data show that the epitopes
recognized by mAbs D33C and CS3 are able to trigger comparable levels
of cell adhesion but are ineffective in stimulating tyrosine
phosphorylation of p125FAK kinase both in suspended and in adherent
cells. In contrast, mAb P9.1.1, which was unable to trigger tyrosine
phosphorylation of p125FAK in suspension, was effective in adhesion
experiments, indicating that the binding site for this antibody elicits
a different response during cell adhesion or integrin clustering in
suspension.
Cell Adhesion to Fibrinogen or to Integrin Antibodies Differently Support Actin Organization
Since p125FAK tyrosine phosphorylation
has been related to the ability of cells to organize focal adhesions
and actin stress fibers, actin cytoskeleton was analyzed in cells
plated for 3 h on IIb
3 antibodies-coated dishes. Cells
plated on mAbs P37, P9.1.1, D33C, and CS3 attached and spread only
partially, but in no case did they organize actin cytoskeleton in
stress fibers (Fig. 5, panels
b-e). A similar pattern of actin organization was found in cells
plated on 10 or 100 µg/ml antibodies, indicating that inability to
trigger cell spreading and actin stress fibers organization was not
dependent on ligand density. Cell adhesion to a 1:1 mixture of the
different antibodies was also unable to induce organization of actin
stress fibers (not shown). Similar organization of
actin cytoskeleton was obtained by
plating cells on mAb 7E2 to hamster
1 subunit or mAb PB1 to the
hamster
5
1 complex (23, 26) (see Figs. 6 and
7). In cells plated on antibodies, actin
was distributed on the periphery of the cells forming filopodia-like structures and large membrane extensions similar to lamellipodia (see
for example, Fig. 5, panel b, and Fig. 6, panel
b). In contrast, the cells spread completely on fibrinogen, and
actin was organized in stress fibers (Fig. 5, panel a).
Treatment of cells with cytochalasin D, an inhibitor of actin
polymerization, completely abolished actin organization, either in
filopodia- or lamellipodia-like structures on monoclonal antibodies
(data not shown). These data show that antibody occupation of specific
epitopes of the
IIb
3 integrin is able to induce actin filopodia-
and lamellipodia-like structures but not full cell spreading nor stress
fibers, as seen when cells bind to purified fibrinogen. Formation of
filopodia- and lamellipodia-like structures was observed both on
anti-
IIb
3 antibodies able to trigger p125FAK tyrosine
phosphorylation (P37, P9.1.1., P4.1.1, P8.2.1, mAb B212) than on
antibodies incapable of inducing p125FAK tyrosine phosphorylation (mAbs
D33C and CS3), indicating that phosphorylation of p125FAK was not
necessary to the actin organization reached by cells plated on
anti-
IIb
3 antibodies. Moreover, since plating cells on
anti-
IIb
3 mAbs able to trigger p125FAK tyrosine phosphorylation
did not lead to the actin stress fibers organization seen on
fibrinogen, these data also indicate that p125FAK phosphorylation is
not sufficient to the assembly of actin stress fibers.
Cell Spreading and Actin Stress Fibers Formation in Cells Plated on Anti-integrin Antibodies: Involvment of PKC and Rho GTPase
Protein kinase C has been implicated in the ability of
cells to spread and to organize focal adhesions on matrix proteins (31,
32). To test whether the inability of cells plated on anti-integrin
antibodies to spread completely and to organize actin stress fibers
could be overcome by activation of the PKC pathway, cells were plated
on mAb B212 to the 3 subunit or mAb 7E2 to the hamster
1 subunit
in the presence of PMA, a known activator of PKC. PMA-treated cells
appeared well spread within 1 h of adhesion (Fig. 6, panels
c and d), whereas nontreated cells remain roughly round
or only partially spread (Fig. 6, panels a and
b). PMA induced spreading at 50 nM, and maximal
extent of cells spreading (evaluated as 90% of the cells becoming
spread) was obtained at 100 nM. PMA-treated cells acquired
a very flat morphology, and the majority of the cell population lost
the filopodia-like extensions that characterized the untreated cells.
At higher PMA concentrations, 500 nM, however, the
peripheral cell edges start to detach from the substratum, suggesting
formation of membrane ruffling (not shown). PMA treatment only promoted
formation of small bundles of actin filaments in 30-50% of the cells,
predominantly located at the periphery of the cell. Otherwise actin
remained organized in small patches in the cytoplasm or at the boundary of the cell (Fig. 6, panels c and d). To further
investigate whether the effects observed with PMA were indeed due to
PKC activation, cells were plated on mAb B212 in the presence of a
selective PKC inhibitor, the mesylate salt RO 31-8220 (33). RO
31-8220-treated cells in the presence of PMA remained round and poorly
attached, showing that the inhibitor prevented PMA-induced full
spreading (Fig. 6, panel e). Moreover, cells also treated
with RO 31-8220 and plated on anti-
IIb
3 antibodies in the
absence of PMA remained round (not shown). RO 31-8220 treatment also
abolishes cell spreading on matrix proteins, fibrinogen, or vitronectin
(not shown). Similar results were obtained by treating cells with
calphostin C, another inhibitor of PKC (not shown), suggesting that
basal activation of PKC may occur in cells adherent to
antibodies-coated dishes and may be involved in organization of
filopodia- and lamellipodia-like structures.
Previous work showed that genistein-sensitive tyrosine kinases are implicated in integrin-mediated cell adhesion (12). Cells plated on anti-integrin antibodies in the presence of PMA and genistein did not spread completely (Fig. 6, panel f), indicating that tyrosine kinase activity is implicated in PMA-induced cell spreading.
Since PMA-mediated PKC activation did not appear to be sufficient for
full organization of actin stress fibers, it is possible that this
event requires the small GTPase Rho (17, 34, 35). To investigate a role
for Rho in actin cytoskeleton organization following integrin antibody
binding, we used the bacterial toxin Escherichia coli CNF1,
known to activate Rho (26, 36-38). As preliminary experiments,
transfected cells were treated with 106 M
CNF1 and plated on vitronectin. CNF1 treatment increases the number of
stress fibers in 100% of the cells, showing a more densely packed
organization of stress fibers than in untreated cells (Fig. 7,
panels a and b).
To investigate whether CNF1-mediated constitutive activation of Rho
might affect actin stress fibers formation on anti-IIb
3 antibodies, cells treated with 10
6 M CNF1
were plated on anti-
IIb
3 antibodies. The cells attached poorly
and remained completely round (Fig. 7, panel c). Prolonged times of adhesion (from 3 to 20 h) or lower doses of CNF1 (from 10
7 to 10
10 M) did not lead to
cell spreading on antibodies and actin stress fibers organization (not
shown). This rather surprising result indicates that activation of Rho
per se does not support actin stress fibers organization
when integrins are occupied by antibodies. Different results were
obtained when cells were treated with the combination of PMA and CNF1.
In this condition, after 1 h of adhesion, cells organize a
prominent array of actin stress fibers on anti-
IIb
3 or
5
1
integrins antibodies (Fig. 7, panels h and i),
indicating that both PKC and Rho activation are necessary to reach
actin stress fibers organization.
To test whether PKC and Rho activation might modify
p125FAK tyrosine phosphorylation, PMA- or PMA and CNF1- treated cells were plated on mAb D33C, which is not able to trigger p125FAK tyrosine
phosphorylation (see also Fig. 4). Cells plated on mAb D33C in the
presence of PMA show induction of p125FAK tyrosine phosphorylation
(Fig. 8). The relevance of tyrosine
kinase activity in PMA-induced cell spreading was also shown by the
fact that genistein, a known inhibitor of tyrosine kinases, blocks
PMA-induced cell spreading (see Fig. 6, panel f). Moreover,
combined treatment with CNF1 and PMA caused a slight additional
increase in the level of p125FAK tyrosine phosphorylation. These data
indicate that p125FAK phosphorylation can be induced by activation of
PKC and Rho pathways.
In this study, we transfected the human IIb
3 integrin in CHO
cells, and we used a panel of monoclonal antibodies directed either to
the human
IIb
3 or to the endogenous hamster
5
1 integrin complexes to dissect the signaling pathways involved in actin organization. We found that while most of the antibodies induce p125FAK
tyrosine phosphorylation, some of them were ineffective. All the
antibodies, however, support adhesion and partial actin organization
but not full spreading and stress fibers polymerization. In contrast,
cells spread and organize stress fibers on fibrinogen. We found that
the combined action of PKC activator PMA and Rho activator CNF1 is
required to trigger stress fibers organization following adhesion to
anti-integrin antibodies. These results allow us to draw the following
conclusions. 1) Different integrin ligands trigger different actin
organization; matrix proteins, such as fibrinogen, are able to induce
actin stress fibers formation, while antibodies to integrins induce
filopodia and lamellipodia but not stress fibers. 2) Actin organization
in filopodia- and lamellipodia-like structures does not require p125FAK
tyrosine phosphorylation. 3) Tyrosine phosphorylation of p125FAK in
response to integrin occupancy is not sufficient per se to
lead to stress fibers organization. 4) To obtain
integrin-dependent full cell spreading and actin stress
fibers organization, at least two events are required, activation of
PKC and Rho.
In this work, we found that the phosphorylation of p125FAK kinase can
be driven only by antibody binding to specific epitopes of the
IIb
3 integrin. The antibodies described in Table I bind to
distinct epitopes of the receptor, some of them (mAbs P9.1.1, P8.2.1,
P4.1.1) inhibit ligand binding (24, 25), while others (D33C, CS3)
induce fibrinogen binding mimicking
IIb
3 activation via ADP (22).
The ability to trigger p125FAK tyrosine phosphorylation does not depend
on the ability of the antibodies to block ligand binding. On the other
hand, two of the antibodies that were ineffective at inducing p125FAK
tyrosine phosphorylation, D33C and CS3, are activating antibodies (22).
The inability of these mAbs to trigger p125FAK tyrosine phosphorylation
was observed both in clustering in suspension and in adhesion to
antibodies-coated dishes. Similar data were reported by Pelletier
et al. (39). Shattil and co-workers showed that an antibody
with properties comparable with D33C and CS3 induces p125FAK tyrosine
phosphorylation in platelets only in the presence of a second stimulus
(40). This is similar to our data showing that PMA and CNF1 are
necessary to obtain p125FAK tyrosine phosphorylation in cells plated on
mAb D33C. In addition, however, we show that using different
antibodies, clustering of the
IIb
3 is sufficient to trigger
p125FAK tyrosine phosphorylation in CHO cells expressing this receptor.
Phosphorylation of the p125FAK in
IIb
3-transfected cells has also
been very recently described (41). These data and the present work
indicate clearly that the
IIb
3 integrin is able to trigger the
cytoplasmic pathway leading to tyrosine phosphorylation of the p125FAK
kinase in response to integrin ligand binding. The requirement of a
co-stimulus in
IIb
3-induced p125FAK tyrosine phosphorylation in
platelets may be due either to the properties of the specific
activating antibody used or to a platelet-specific signaling
mechanism.
The data presented here indicate that antibodies to either 5
1 or
IIb
3 integrins are able to support cell adhesion, but they are
incapable of triggering stress fibers formation. This experimental
system allows us to test biological pathways implicated in
integrin-mediated cells adhesion and cytoskeletal organization. Cells
plated on the different antibodies were poorly spread, and F-actin was
organized in small patches and in peripheral thin filopodia-like
structures. Similar structures have been previously described in the
early phases of cell adhesion, as a peculiar actin organization during
focal adhesion formation, in a step which precedees accumulation of
cytoskeletal elements (42, 43). Assembly of short bundles of actin
filaments to produce microspikes and filopodia have also been
previously demonstrated in motile cells and at the end of neuron growth
cones (44). The formation of filopodia and lamellipodia is induced by
Cdc42 and Rac, belonging to the Rho family of small GTP-binding
proteins (17, 45). We can hypothesize that interaction with monoclonal
antibodies leads to integrin-dependent activation of Cdc42 and
Rac. Thus, cell adhesion to the monoclonal antibodies to integrin
receptors leads to an initial level of actin organization, but it is
insufficient to drive redistribution of cytoskeletal components to
organize actin in stress fibers.
We found that concomitant treatment with PMA and CNF1, activators of
PKC and Rho respectively, is required to obtain cell spreading and
actin stress fibers organization in cells plated on anti-integrin
antibodies. These pathways need to be activated in cells plated on
antibodies directed either to 1 or
3 integrin subunits,
indicating that both
1- or
3-mediated integrin clustering requires similar intracellular pathways to trigger actin stress fibers
formation.
PKC and Rho appear to play different roles in controlling these processes, while PKC controls cell spreading and Rho in combination with PKC regulates organization of stress fibers. In the presence of the phorbol ester PMA, filopodia disappear, and the cells become round, well spread with large lamellipodia. Short and thin bundles of actin filaments, however, were visible, located at the periphery of the cells, and actin was mainly organized in small patches inside the cytoplasm. Since this effect was blocked by the PKC inhibitors RO 31-8220 and calphostin C, these data strongly suggest that cell spreading on integrin antibodies is regulated by PKC. PKC involvment in spreading and focal adhesion formation has been demonstrated in cells plated on matrix proteins (31, 32, 46). Using specific inhibitors or activators, it has been shown that PKC regulates spreading of platelets on fibrinogen (46) and of fibroblasts on fibronectin (32) or the ability of fibroblasts to organize focal adhesions on fibronectin (31). It has been recently reported that syndecan 4, a proteoglycan localized in the focal adhesions in a PKC dependent manner, can activate PKC and thus may function as a co-receptor with integrins, interacting with the heparin binding domain exposed on the fibronectin molecule (47).
Our results show that stress fibers formation requires additional treatment with the Rho activator CNF1. The action of CNF1 is specific for Rho since this toxin activates Rho by specifically deamidating glutamine at position 63 (38). Interestingly in our system, we can show that Rho, although essential, is not sufficient to drive actin stress fibers formation. In fact, cells treated with CNF1 remained round and did not organize stress fibers on antibodies-coated dishes. A dose-response experiment, where cells were treated with decreasing doses of CNF1, did not allow us to find a condition in which CNF1-treated cells were able to spread on anti-integrin antibodies. Since activation of Rho has been shown to stimulate cell contractility (48), it is possible that in cells where Rho has been constitutively activated by CNF1, cell contractility may antagonize cell spreading. In contrast, cells plated on vitronectin respond to CNF1 by organizing thick and prominent stress fibers, further showing that matrix proteins and integrin antibodies elicit distinct intracellular signaling responses.
The data reported in this work suggest that, in addition to PKC and Rho
activation, tyrosine kinases are likely to be important in regulation
of stress fibers formation. The results presented here support the idea
that tyrosine phosphorylation of p125FAK per se is not
sufficient to induce organization of stress fibers since cells adhering
to antibodies able to trigger p125FAK tyrosine phosphorylation do not
organize actin stress fibers. PKC and Rho activation, however, induces
tyrosine phosphorylation of p125FAK kinase in cells plated on
antibodies that are not able per se to activate this
phosphorylation event. According to these data, we and others have
previously shown that protein tyrosine kinase (PTK) inhibitors blocked
formation of stress fibers (10, 12), thus suggesting that PTKs are
required in addition to PKC and Rho to achieve organization of actin
stress fibers in response to integrin ligand binding
(Fig. 9). Contrasting results on the role
of p125FAK tyrosine phosphorylation in stress fibers and focal adhesion
assembly have been presented (49). Cells deriving from p125FAK
knock-out mice assemble focal adhesions in the absence of p125FAK but
reduce their motility (50). Gilmore and Romer (51), using a dominant
negative form of p125FAK, found a decreased cell motility but not
interference on focal adhesion organization. These results suggest that
p125FAK or its tyrosine phosphorylation, rather than playing a
structural role, may be crucial in regulating turn-over of stress
fibers and focal adhesions.
Involvment of PKC and Rho in p125FAK tyrosine phosphorylation is in agreement with findings showing that PMA can trigger p125FAK tyrosine phosphorylation in cells plated on matrix proteins (32, 46). The ability of Rho to trigger p125FAK tyrosine phosphorylation has been recently reported, following scraping of activated Rho into cells (52).
The data presented above indicate that p125FAK tyrosine phosphorylation is not able to trigger PKC or Rho activation since antibodies capable of inducing p125FAK phosphorylation did not lead to cell spreading or actin stress fibers, processes which require PKC and Rho pathways. At the same time, PKC does not trigger Rho activation and vice versa, as shown by the fact that stimulation by PMA only leads to cell spreading but not to stress fibers organization. CNF1-mediated Rho activation is also unable to drive actin cytoskeleton organization to stress fibers. Thus p125FAK tyrosine phosphorylation and PKC and Rho activation are three independent pathways, which are all required to trigger full cell spreading and actin stress fibers. In our system, anti-integrin antibodies are able to trigger only the pathway leading to p125FAK tyrosine phosphorylation, while fibrinogen is likely to activate all three pathways.
Our data show that organization of filopodia-like structures does not
require induction of p125FAK tyrosine phosphorylation by IIb
3. In
fact, the
IIb
3-dependent organization of
filopodia-like structures occurs also with antibodies that are
incapable of activating this signaling event. Moreover, the ability of
the
IIb
3 integrin to organize F-actin independently from p125FAK
tyrosine phosphorylation was demonstrated by clustering of the
IIb
3 integrin on the surface of cells in suspension. In these
cells, in fact, clustering with an antibody incapable of triggering
p125FAK tyrosine phosphorylation, leads to actin co-clustering. Thus,
actin aggregation under the clusterized receptor or formation of
filopodia-like structures in adhesion to antibodies-coated dishes are
events independent from p125FAK tyrosine phosphorylation. Whether other
tyrosine kinases are involved is not known at present. Tyrosine kinase activity is required for accumulation of F-actin and other focal adhesion proteins in the clustering site (53). The data presented here
indicate that p125FAK tyrosine phosphorylation is not involved in this
step.
In conclusion, the data reported in this paper indicate that PKC and Rho are required in cell spreading and actin stress fibers assembly following integrin-dependent signaling. Although PKC involvment in spreading or focal adhesion formation on fibronectin has been previously shown (31, 32) and Rho activation has been demonstrated to regulate actin stress fibers formation following mitogen stimulation in mouse fibroblasts (35, 45), the data presented here are the first demonstration in a single cell line that following integrin-ligand binding, PKC and Rho are both necessary to the establishment of cell spreading and actin stress fibers organization.
We are grateful to Drs. J. Gonzales-Rodriguez and Zaverio Ruggeri for the generous gift of antibodies. We also thank Dr. S. Shattil for helpful comments on the manuscript.