(Received for publication, July 2, 1996)
From the Department of Hematology and Oncology,
Clinical Sciences for Pathological Organs, Graduate School of Medicine,
Kyoto University, 54 Shogoin-Kawaramachi, Sakyo-ku, Kyoto 606-01, ¶ Department of Clinical Pathology, Tokyo Medical College, Tokyo
160,
Department of Cardiovascular Research, Tokyo Metropolitan
Institute of Medical Science, Tokyo 113, and ** Blood Transfusion
Division, Niigata University Medical Hospital, Niigata 951, Japan
Activation of circulating platelets by
subendothelial collagen is an essential event in vascular hemostasis.
In human platelets, two membrane glycoprotein (GP) abnormalities,
integrin 2
1 deficiency and GPVI
deficiency, have been reported to result in severe hyporesponsiveness to fibrillar collagen. Although it has been well established that integrin
2
1, also known as the GPIa-IIa
complex, functions as a primary platelet adhesion receptor for
collagen, the mechanism by which GPVI contributes to collagen-platelet
interaction has been ill defined to date. However, our recent
observation that GPVI cross-linking couples to cyclic AMP-insensitive
activation of c-Src and Syk tyrosine kinases suggested a potential role
for GPVI in regulating protein-tyrosine phosphorylation by collagen (Ichinohe, T., Takayama, H., Ezumi, Y., Yanagi, S., Yamamura, H., and
Okuma, M. (1995) J. Biol. Chem. 270, 28029-28036). To further investigate this hypothesis, here we examined the
collagen-induced protein-tyrosine phosphorylation in GPVI-deficient
platelets expressing normal amounts of
2
1. In response to collagen, these
platelets exhibited
2
1-dependent c-Src activation
accompanied by tyrosine phosphorylation of several substrates including
cortactin. In contrast, severe defects were observed in
collagen-stimulated Syk activation and tyrosine phosphorylation of
phospholipase C-
2, Vav, and focal adhesion kinase, implicating a
specific requirement of GPVI for recruiting these molecules to
signaling cascades evoked by collagen-platelet interaction.
Upon interaction with subendothelial collagen, platelets undergo a
series of biochemical events that eventually lead to cell spreading,
secretion of granular contents, and activation of major fibrinogen
receptor, integrin IIb
3, to form platelet
aggregates (1). To accomplish this essential role in primary
hemostasis, platelets are believed to express specific surface
receptors for fibrillar collagen. Although a number of candidates have
been proposed as a platelet collagen receptor to date (2), true participation of such molecules in physiological collagen-platelet interaction remains largely unclarified, and the molecular mechanism by
which collagen induces platelet activation has yet to be elucidated. Several lines of evidence have indicated that the integrin heterodimer
2
1 (3, 4), also called glycoprotein
(GP)1 Ia-IIa, is a principal platelet
adhesion receptor for collagen. Clinical deficiency of platelet
2
1 results in a mild to severe bleeding
tendency, accompanied by defective platelet aggregation in response
only to collagen (5, 6, 7). Specific monoclonal antibodies against
2
1 inhibits platelet adhesion to
fibrillar collagen (8, 9), and liposomes containing
2
1 can adhere to collagen types I, II,
III, and IV (10). In addition, the molecular analysis of this complex
revealed the presence of a putative collagen-binding domain within the
2 subunit (11, 12). Although
2
1-mediated adhesion thus appears to be
an essential primary step in collagen-platelet interaction, it is still
unknown whether collagen-
2
1 binding is
sufficient for stimulating a full picture of collagen-induced platelet
activation.
Protein-tyrosine phosphorylation is now considered a key signaling
event in the activation of platelets (13). Collagen is also known to
stimulate activation of protein-tyrosine kinases (PTKs), c-Src (14),
Syk (15), and focal adhesion kinase (Fak) (16) and promotes rapid
tyrosine phosphorylation of a number of proteins in platelets including
cortactin (13), phospholipase C (PLC)-2 (17, 18) and Vav (19).
Unlike the case with other platelet agonists, collagen-stimulated
protein-tyrosine phosphorylation is unique in that it is insensitive to
the inhibition by cAMP-increasing reagents (20), which are known to
potently arrest platelet activation by most agonists. Although integrin
2
1-mediated cell adhesion is believed to
be essential for the induction of such tyrosine phosphorylation (21,
22), another strong candidate responsible for the collagen-stimulated
protein-tyrosine phosphorylation is a so-far uncharacterized 62-kDa
membrane protein, GPVI (23, 24).
Previous clinical reports have suggested an important role of GPVI for
the induction of collagen-stimulated signaling events in platelets.
Despite the normal expression of 2
1
complex, platelets lacking GPVI do not become aggregated by collagen
(23, 24, 25, 26, 27) and exhibit profound defects in collagen-induced platelet
responses such as calcium mobilization (25) and thromboxane synthesis
(23). We demonstrated recently that signaling through GPVI is
considerably dependent on the activation of PTKs in platelets and that
GPVI cross-linking leads to cAMP-insensitive activation of c-Src and
Syk tyrosine kinases (28). In addition, GPVI cross-linking also
stimulates cAMP-insensitive tyrosine phosphorylation of PLC-
2 in a
manner similar to collagen stimulation (28). These observations support
a concept that collagen-stimulated protein-tyrosine phosphorylation in
platelets may be partly regulated through a GPVI-dependent mechanism.
To further test this hypothesis, we chose to examine whether collagen
could induce protein-tyrosine phosphorylation in platelets lacking GPVI
but expressing normal amounts of 2
1
complex. In these platelets, collagen stimulated prompt activation of
c-Src accompanied by tyrosine phosphorylation of a number of proteins in a manner dependent on collagen-
2
1
interaction. However, GPVI-deficient platelets exhibited severe defects
in the activation of Syk and tyrosine phosphorylation of several
substrates by collagen. Our results indicated that collagen-stimulated
protein-tyrosine phosphorylation in platelets are regulated through
both
2
1-dependent and
GPVI-dependent mechanisms and that GPVI-mediated signaling
is required for collagen-stimulated activation of Syk.
Acid-soluble fibrillar collagen prepared from equine tendon was purchased from Horm-Chemie (Munich, Germany). Prostacyclin (PGI2) was kindly provided by Ono Pharmaceutical Co. (Osaka, Japan). Arg-Gly-Asp-Ser (RGDS) tetrapeptide and histone from calf thymus (subgroup f2b) were obtained from Sigma. Phenylmethylsulfonyl fluoride, aprotinin, and sodium orthovanadate was from Nacalai Tesque, Inc. (Kyoto, Japan). All other reagents were obtained as reported previously (28, 29).
AntibodiesMonoclonal antibody (mAb) against c-Src (327)
was obtained from Oncogene Science, Inc. (Uniondale, NY). Anti-Syk mAb
(101) was from Wako Pure Chemical Industries, Ltd. (Osaka, Japan).
Anti-phosphotyrosine mAb (4G10), anti-Fak mAb (2A7), and mAb against
p80/85 cortactin (4F11) were obtained from Upstate Biotechnology, Inc.
(Lake Placid, NY). mAb against human 2
1
(Gi9) was from Immunotech S.A. (Marseille, France), and anti-Fc
RII
mAb (IV.3) was from Medarex, Inc. (Annandale, NJ). Platelet-nonbinding
mouse monoclonal IgG1 (R118) was a generous gift from Dr.
Kina (Chest-Disease Institute, Kyoto University). Rabbit polyclonal IgG
against PLC-
2 or p95Vav was purchased from Santa Cruz Biotechnology,
Inc.(Santa Cruz, CA). Rabbit anti-mouse IgG and peroxidase-conjugated
goat anti-mouse or anti-rabbit IgG were from Cappel Organon Teknika Co.
(Durham, NC).
After informed
consent was obtained, venous blood was collected from healthy adult
donors or three donors with platelet GPVI deficiency reported
previously (23, 26, 27). Washed platelets were prepared as described
previously (28), suspended at 0.5 or 1.0 × 109/ml,
and stimulated with 20 µg/ml of collagen while being stirred at 800 rpm in a glass cuvette. In the experiments to evaluate the effects of
inhibition of collagen-2
1 interaction,
platelet suspensions from GPVI-deficient donors were preincubated for
15 min with 20 µg/ml anti-
2
1 mAb Gi9 to
inhibit collagen-
2
1 interaction, anti-Fc
RII mAb IV.3 as platelet-binding control, or mouse monoclonal IgG1 R118 as platelet-nonbinding control. In some other
experiments, GPVI-deficient platelet suspensions were also pretreated
with several inhibitors for platelet activation including 3 µM PGI2 (3 min), 20 µM
cytochalasin D (10 min), and 1 mM RGDS (5 min).
To prepare whole platelet lysates, each reaction was terminated at the indicated time points by the addition of an equal volume of 2 × SDS sample buffer (28). The samples were boiled for 5 min and subjected to immunoblotting analysis. Immunoprecipitation of each specified protein was performed essentially as described previously (28). Whereas c-Src was immunoprecipitated from lysates prepared in 1% Triton X-100-containing buffer reported by Clark and Brugge (30), the other proteins were precipitated in radioimmunoprecipitation assay (RIPA) buffer (1% Triton X-100, 0.1% SDS, 1% deoxycholate sodium, 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 5 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 20 µg/ml leupeptin, 20 µg/ml aprotinin, and 1 mM sodium orthovanadate). In some experiments, platelet lysates were incubated for 40 min with protein G-agarose (Santa Cruz Biotechnology) prior to the immunoprecipitation to remove preincubating mAbs.
ImmunoblottingImmunoblotting analysis of whole platelet
lysates or immunoprecipitated proteins was performed as described
previously (28). Membranes were probed as indicated with
anti-phosphotyrosine 4G10 (1 µg/ml), anti-Src 327 (1 µg/ml),
anti-Syk 101 (3 µg/ml), anti-Fak 2A7 (5 µg/ml), anti-cortactin 4F11
(5 µg/ml), rabbit anti-PLC-2 (1 µg/ml), or rabbit anti-Vav (1 µg/ml) followed by enhanced chemiluminescence detection of
peroxidase-conjugated secondary antibody binding.
In vitro kinase activity of c-Src or Syk was measured as described previously (28). Histone phosphorylating activity was determined by the ratio of incorporation of 32P quantitated by a BAS 2000 imaging analyzer (Fuji) to the relative amount of immunoprecipitated c-Src or Syk estimated by densitometric analysis of quantitative immunoblots corresponding to each sample.
GPVI-deficient platelets were obtained from three
subjects reported previously (23, 26, 27). Immunoblotting analyses confirmed that GPVI was undetectable in one of these patients (donor 1)
(23) and less than 10% normal in the other two (donors 2 and 3) (26,
27). These platelets subnormally adhered to fibrillar collagen (types I
and III) but specifically lacked collagen-induced aggregation despite
expressing normal amounts of cell surface 2
1 complex (23, 26, 27). Because
protein-tyrosine phosphorylation is suggested to be one of the earliest
platelet responses to collagen (22), we first examined whether the
interaction with collagen promotes tyrosine phosphorylation in
GPVI-deficient platelets. Although previous reports have demonstrated
that GPVI-deficient platelets were almost totally defective in
collagen-stimulated signaling events (23, 25), a substantial degree of
rapid protein-tyrosine phosphorylation was induced in response to
collagen in the platelets from all three donors we examined (Fig.
1A). Although no apparent difference was
found in a profile of major tyrosine-phosphorylated protein bands (150, 130, a bundle of 68-72, and 64 kDa) between normal and GPVI-deficient
platelets, time-dependent dephosphorylation of these bands
was not observed in the GPVI-deficient platelets (Fig. 1B),
which could be explained by the lack of
IIb
3-mediated signaling required for the
regulation of proteintyrosine phosphatases as we have demonstrated
previously (31, 32).
Integrin 2
1 is believed to mediate
primary platelet adhesion to collagen (4) and is also implicated in the
induction of protein-tyrosine phosphorylation by collagen (21, 22). Confirming this, pretreatment of GPVI-deficient platelets with anti-
2
1 mAb completely abolished
collagen-induced tyrosine phosphorylation, whereas the preincubation
with either platelet-binding control antibody (anti-Fc receptor mAb) or
platelet-nonbinding mouse IgG1 (Fig.
2A) did not. In addition, cytochalasin D, an
inhibitor of F-actin formation, also significantly abrogated the
induction of such protein-tyrosine phosphorylation, whereas the
presence of PGI2, which most potently inhibits platelet
activation by elevating intracellular cAMP, or RGDS tetrapeptide, which
blocks ligand-
IIb
3 binding, was without
effect (Fig. 2B). These findings indicated that collagen
stimulates protein-tyrosine phosphorylation even in the absence of GPVI
in a manner dependent on both collagen-
2
1 interaction and subsequent cytoskeletal rearrangement.
Collagen Does Not Stimulate Integrin-mediated Tyrosine Phosphorylation of Fak in GPVI-deficient Platelets
In normal
platelets, collagen has been reported to promote activation of c-Src
(14), Syk (15), and Fak (16) tyrosine kinases. Although the mechanism
of how collagen-platelet interaction leads to activation of these PTKs
is not fully clarified, previous analysis suggested the involvement of
integrin 2
1 in collagen-induced activation of Fak (21). However, under our experimental conditions, it
appeared unlikely that Fak becomes initially activated solely by
collagen-
2
1 interaction because tyrosine
phosphorylation of Fak occurred later than peak activation of c-Src and
Syk in normal platelets (Fig. 3A and Fig.
4A). Furthermore, collagen did not induce
tyrosine phosphorylation of Fak in GPVI-deficient platelets (Fig.
4A), suggesting that Fak is not involved in the
2
1-mediated tyrosine phosphorylation
observed in the GPVI-deficient platelets. Because collagen does not
stimulate significant calcium mobilization in GPVI-deficient platelets
(25), these findings may be consistent with the hypothesis suggested by
Shattil et al. (33) that tyrosine phosphorylation of Fak
requires both ligand-integrin interaction and signaling from other
receptors leading to either calcium mobilization or protein kinase C
activation.
Collagen Rapidly Stimulates Activation of c-Src but not of Syk in GPVI-deficient Platelets
To investigate responsible PTKs for the
collagen-induced tyrosine phosphorylation in GPVI-deficient platelets,
collagen-stimulated activities of c-Src and Syk were next evaluated by
their in vitro histone phosphorylating activity (Fig. 3,
A and B). In response to collagen, the activity
of c-Src from GPVI-deficient platelets exhibited a 1.5-fold increase
that was comparable to normal platelets. Such activation of c-Src was
found to be also 2
1-dependent
because it was significantly inhibited in the presence of
anti-
2
1 mAb (Fig. 3C).
By sharp contrast with such rapid activation of c-Src, collagen-stimulated activation of Syk was invariably impaired in GPVI-deficient platelets. In response to collagen, Syk in normal platelets became rapidly phosphorylated on tyrosine and exhibited a 5-fold maximal increase in its activity. However, in the GPVI-deficient platelets, an increase in collagen-stimulated Syk activity was markedly blunted, and tyrosine phosphorylation of Syk was undetectable (Fig. 3, A and B). This defect in Syk activation in GPVI deficiency was specific to stimulation with collagen, because other platelet agonists such as thrombin and STA2 (an agonistic analogue of thromboxane A2) induced prompt activation and subsequent deactivation of Syk in GPVI-deficient platelets in a profile indistinguishable from that of normal platelets (data not shown).
Collagen-stimulated Tyrosine Phosphorylation of PLC-Because a number of
platelet agonists usually stimulate parallel activation of c-Src and
Syk in normal platelets, it was highly implicative to us that c-Src was
preferentially activated in collagen-treated GPVI-deficient platelets.
To gain some insight into the subsequent cellular events, we examined
the difference in substrates that become tyrosine-phosphorylated in
response to collagen between normal and GPVI-deficient platelets (Fig.
4B). In collagen-stimulated normal platelets, we could
readily detect tyrosine phosphorylation of cortactin, an F-actin
binding protein originally identified in v-Src-transformed cells (34,
35), PLC-2 that may be involved in collagen-stimulated calcium
mobilization (17), and Vav, an adapter oncoprotein that may participate
in Syk-dependent signaling (36, 37). Tyrosine
phosphorylation of these substrates by collagen was independent of
platelet aggregation, because it was still observed in the presence of
inhibitors for platelet aggregation such as RGDS peptide or
PGI2 (data not shown). In GPVI-deficient platelets,
collagen also induced tyrosine phosphorylation of cortactin in an
2
1-dependent manner (Fig. 4,
B and C) but failed to stimulate that of PLC-
2
and Vav (Fig. 4B). On the other hand, in normal platelets,
antibody-mediated GPVI cross-linking is sufficient to stimulate
cAMP-insensitive tyrosine phosphorylation of Syk, PLC-
2 (28), and
Vav (data not shown). Very recently, Cichowski et al. (19)
reported that collagen-stimulated tyrosine phosphorylation of Vav is an
2
1-dependent event, but our
above findings suggested an additional requirement of GPVI-mediated
signaling for Vav to become phosphorylated on tyrosine.
Thus far, several clinical reports have suggested the involvement
of GPVI in the events responsible for collagen-induced platelet aggregation, but its precise role in collagen-platelet interaction has
been poorly defined. Although the lack of collagen-stimulated responses
in 2
1-deficient platelets is considered
to be a consequence of their defective adhesion to collagen (2, 4),
previous studies could not explain the mechanism of how GPVI-deficient platelets with normal amounts of
2
1
became severely hyporesponsive to collagen. In this study, we
demonstrated for the first time that collagen was capable of
stimulating protein-tyrosine phosphorylation in GPVI-deficient
platelets that lack most other signaling events evoked by collagen. In
addition, we showed that such tyrosine phosphorylation was almost
totally diminished in the presence of
anti-
2
1 mAb, indicating that
collagen-
2
1 interaction is substantially
retained to stimulate protein-tyrosine phosphorylation in the absence
of GPVI.
Among three PTKs that have been reported to become activated by
collagen in normal platelets (14, 15, 16), we could detect increased
activity of only c-Src in collagen-stimulated GPVI-deficient platelets.
Such activation of c-Src that accompanied tyrosine phosphorylation of
its putative substrate, cortactin (35), was also inhibited when
GPVI-deficient platelets were pretreated with
anti-2
1 mAb. These results indicated that
integrin
2
1-mediated events were
responsible for c-Src-dependent protein-tyrosine phosphorylation when GPVI-deficient platelets reacted with collagen. By
contrast, GPVI-deficient platelets did not exhibit detectable tyrosine
phosphorylation of Syk, Fak, PLC-
2, and Vav in response to collagen.
Therefore, although it has been believed that protein-tyrosine phosphorylation in collagen-activated platelets is mostly dependent on
collagen-
2
1 interaction (21), our present
observations indicated an additional requirement of GPVI as a signaling
counterpart to complement
2
1-mediated
signalings for the collagen-stimulated tyrosine phosphorylation of
certain specific substrates. Furthermore, the significant abolishment
of collagen-induced c-Src activation in
anti-
2
1 mAb-treated, GPVI-deficient
platelets strongly suggests that collagen-stimulated activation of
c-Src and Syk in platelets are, for the most part, regulated through
2
1 and GPVI.
Collagen-platelet interaction has been considered to be a complex
process involving
2
1-dependent and
2
1-independent mechanisms (2, 38, 39).
Although whether GPVI can mediate
2
1-independent platelet interaction with
collagen is still unclarified, a recent clinical report (7) describing
that
2
1-deficient platelets expressing
normal amounts of GPVI did not respond to collagen raised a question as
to the role of GPVI as an independent signaling receptor for collagen.
When we tested the effects of anti-
2
1 mAb
on collagen-stimulated activation of c-Src and Syk in normal platelets,
this treatment almost totally abolished such events under conditions of
stasis.2 However, as partly shown in a
previous report (40), we observed delayed but significant activation of
both kinases by collagen when anti-
2
1
mAb-treated normal platelets were stirred,2 suggesting that
anti-
2
1-mediated inhibition of
collagen-
2
1 interaction is insufficient
for totally preventing collagen-induced activation of normal platelets,
especially under dynamic conditions. Although the involvement of
GPVI-mediated signaling under such conditions has yet to be elucidated,
GPVI may cooperate with
2
1-mediated signaling in a manner dependent on the extent of flow conditions where
platelets react with collagen.
Another candidate regulator for collagen-stimulated protein-tyrosine
phosphorylation in platelets is GPIV (CD36) (41), which is shown to
associate with Src family tyrosine kinases Fyn, Lyn, and Yes (42).
However, in contrast to the 2
1-deficient
or GPVI-deficient platelets, platelets lacking GPIV became normally aggregated (43) and exhibited essentially normal protein-tyrosine phosphorylation in response to collagen (44). Therefore, although GPIV
has been suggested to be an important mediator of platelet adhesion to
collagen (39, 45), its role in collagen-induced signaling events seems
to be relatively minor when compared with those of
2
1 or GPVI.
According to our present study,
2
1-mediated activation of c-Src and
tyrosine phosphorylation of cortactin did not link with either
activation of Syk or tyrosine phosphorylation of PLC-
2 and Vav in
the absence of GPVI. Given the observations that the earliest known
event following GPVI engagement is activation of c-Src and Syk tyrosine
kinases (28) and that collagen is capable of activating c-Src also via
an
2
1-dependent mechanism as
shown in this study, GPVI appears to possess a unique function unshared with
2
1 in regulating tyrosine
phosphorylation and activation of Syk. If this is the case, the lack of
Syk-dependent signalings in response to collagen should be
the most reasonable explanation for a distinctive molecular pathology
leading to defective collagen-induced aggregation in GPVI-deficient
platelets. Of note here, the essential requirement of
Syk-dependent signalings for PLC-mediated calcium mobilization and subsequent cell activation was demonstrated in other
cells pharmacologically or genetically engineered to compromise Syk
tyrosine kinase activity (46, 47). In this sense, GPVI deficiency could
be regarded as the first clinical example for an impaired cellular
function due to uncoupling of ligand-cell interaction with the
activation of Syk family PTKs.
In summary, we have described altered collagen-stimulated
protein-tyrosine phosphorylation in GPVI-deficient platelets where 2
1-dependent signalings were
substantially retained. Although GPVI engagement leads to activation of
both c-Src and Syk in platelets, collagen-stimulated activation of
c-Src is also independently shared by
2
1-mediated events in contrast to the Syk
activation, which seems to exclusively require the presence of GPVI.
Fig. 5 summarizes our working hypothesis for the
differential roles of
2
1 and GPVI in
collagen-stimulated protein-tyrosine phosphorylation in platelets.
We thank Dr. Tatsuo Kina for generously providing us platelet nonbinding mouse monoclonal IgG1 R118; Dr. Kenjiro Tanoue for valuable discussions; Drs. Kenjiro Tomo and Kenzo Hirai for encouragement and suggestions; and Hiroko Nakagawa and Ikuko Nakamura for excellent technical and secretarial assistance.