(Received for publication, September 13, 1996, and in revised form, December 27, 1996)
From the Life Science Research Center, Advanced Technology Research Laboratories, Nippon Steel Corporation, 1618 Ida Nakahara-ku, Kawasaki 211, Japan
To develop a specific antagonist of platelet
IIb
3 using small linear peptides, we synthesized a series of
hexapeptides that did not have an Arg-Gly-Asp (RGD) sequence and
examined their anti-platelet activity and their specificity for
IIb
3. We found a novel motif sequence,
Pro-X1-X2-X3-Asp-X4,
where X1 to X4 were all L-form
-amino acids,
which specifically inhibited aggregation of human platelets at
submicromolar concentrations. The Pro residue at the N terminus was
essential to the anti-platelet activity, and the acetylation of the
imino group of this residue also resulted in the complete loss of the
activity. The results of the binding assay using purified human
platelet
IIb
3 and placental
v
3 and those of the cell
adhesion assay suggest that this motif peptide is highly specific for
platelet
IIb
3 among other integrins. Flow cytometric studies
using an fluorescein isothiocyanate-labeled RGD peptide showed that
this motif peptide inhibited the binding of an RGD peptide to activated
platelets, suggesting that it has the same inhibitory mode as RGD
peptides. Conformational analysis of this motif peptide and an
RGD-containing peptide suggests that the imino group of the Pro residue
may substitute for the role of the guanidino group of the Arg residue
of the RGD sequence.
Integrins are heterodimeric cell surface receptor molecules and
are thought to be particularly important mediators of cell adhesion,
cell migration, and adhesion-dependent intracellular signaling. Until now, more than 20 integrin receptors have been identified, and many of them recognize the Arg-Gly-Asp
(RGD)1 sequence as a common recognition
motif within their putative ligands upon ligand-receptor interaction
(1-4). The ligand binding specificity has been studied by many
investigators, and it has been revealed that some integrins show
apparent redundancy in integrin-ligand interactions. For example,
v
3, a so-called vitronectin receptor, can bind several apparently
dissimilar ligands such as vitronectin, fibrinogen, fibronectin, von
Willebrand factor, osteopontin, and bone sialoprotein, all of which
have the RGD sequence (reviewed in Refs. 5 and 6). In addition, the
same RGD site within a ligand can be recognized by several different integrins. These apparent redundancies in integrin-ligand interactions are very interesting aspects of integrin receptors and may be beneficial to the cell adhesion of many types of cells to extracellular matrix proteins.
We are interested in developing small peptides that are specific for
each integrin as a tool for investigating ligand-integrin interactions
and as an agent for therapeutic uses. Recently, many small peptides
containing the RGD sequence have been synthesized, and their
specificity toward integrins has been examined (7, 8). Most of the
linear RGD peptides have shown very low specificity among many types of
integrin receptors including platelet IIb
3 (fibrinogen receptor),
v
3 (vitronectin receptor), and
5
1 (fibronectin receptor).
In the case of
IIb
3 and
v
3, both integrins have the same
subunit (
3), although they have distinct
subunits. It has
been reported that the primary amino sequences of
IIb and
v are
only about 36% homologous (9). By cross-linking studies, it has been
shown that RGD peptides were bound and cross-linked with the
3
subunit (10, 11). Moreover, single point mutation at several positions
within the
3 subunit abrogated the binding of RGD peptides to both
receptors (12-15). These data indicate that the interaction between
the RGD sequence and the common
3 subunit is responsible for the
ligand recognition by both integrins. However, it has also been
reported that some RGD peptides with a cyclic structure show relatively
high specificity (16-18). These results suggest that the
three-dimensional structure of the RGD site itself and/or its flanking
regions are very important for the ligand binding specificity, because
cyclic peptides have very rigid structure and their three-dimensional
structures are considerably fixed. It is also suggested that the
subunit, the critical difference between both integrins, is associated
with such a ligand binding specificity. This idea is supported by the
results of experiments in which a decapeptide from the carboxyl
terminus of the fibrinogen
-chain was relatively specific for
platelet
IIb
3 and it cross-linked with
IIb subunit of this
integrin (19, 20). Several peptides without the RGD sequence, which are
highly specific for platelet
IIb
3, have also been reported. For
example, a tick saliva protein, which does not have the RGD sequence,
is a potent
IIb
3 antagonist and shows high specificity for this
integrin (21). Another example is barbourin, which is a member of the
disintegrins found in snake venom (22).
The aim of our study is to develop a potent and highly specific
IIb
3 antagonist of small linear peptides (less than six residues)
consisting of natural amino acids. We found a novel motif sequence of a
hexapeptide. One of the motif peptides, Pro-Ser-Hyp-Gly-Asp-Trp (Hyp
representing 4-hydroxy-L-proline) (NSL-9511), inhibited
platelet aggregation at a submicromolar concentration, although it did not affect adhesion of human endothelial cells to immobilized vitronectin and fibronectin. The results of the solid phase binding assay suggest that this peptide is a specific antagonist of platelet
IIb
3. NSL-9511 also exhibited antithrombotic activity in an in vivo thrombosis model.
Platelet aggregation studies were performed in platelet-rich plasma (PRP) obtained from various animal species, including humans. Blood was drawn into plastic syringes containing 1/10 volume of 3.8% trisodium citrate. PRP was prepared by centrifugation of citrated whole blood at 160 × g for 15 min at room temperature. PRP was removed, and the platelet count was determined. Platelet-poor plasma was obtained by centrifugation of the remaining blood at 2000 × g for 15 min. Saline or peptide solution of various concentrations was added to PRP at 37 °C 1 min prior to the initiation of platelet aggregation. Platelet aggregation was initiated with 5 µg/ml collagen, and the aggregation was measured in an aggregometer (NBS Hematracer-601, Nikoh Bioscience Co., Ltd., Tokyo) as an increase in light transmission. Platelet aggregation is presented as the percentage of inhibition of the rate of platelet aggregation compared with control samples, and the IC50 values were calculated from dose-inhibition curves.
Preparation of Integrin ReceptorsVitronectin receptor
(v
3) was purified according to the procedures described elsewhere
(17, 23) with some modifications. Briefly, human placentas were cut
into small pieces and extracted with buffer containing 100 mM octyl glucoside. The extract was centrifugated, and the
supernatant was applied to a Sepharose CL4B column and then applied to
a GRGDSPK-Sepharose affinity column. The column was washed with five
bed volumes of Tris-buffered saline (TBS) containing 3 mM
Ca2+ and 50 mM octyl glucoside at 4 °C and
then washed with five bed volumes of the same solution at room
temperature. Bound vitronectin receptor was eluted with TBS containing
10 mM EDTA and 50 mM octyl glucoside. Finally,
Ca2+ was added to the eluted fraction so that the final
concentration of the free Ca2+ was 2 mM.
Platelet fibrinogen receptor (IIb
3) was purified from outdated
human platelets using a procedure almost identical to that for
v
3. Briefly, the platelets were washed three times with TBS
containing 1 mM EDTA and 0.2% glucose and then were lysed with TBS containing 100 mM octyl glucoside, 1 mM MnCl2, 1 mM MgCl2, and 0.1 mM phenylmethylsulfonyl fluoride. After the
centrifugation at 30,000 × g, the supernatant fraction
was applied to a Sepharose 2B column, and the flow-through fraction was
then applied to a GRGDSPK-Sepharose affinity column. Bound fibrinogen
receptor was eluted with TBS containing 50 mM octyl
glucoside and 5 mM GRGDS peptide. The eluted fraction was
dialyzed to remove a GRGDS peptide and concentrated on an Amicon YM 30 filter. Purified human
IIb
3, purchased from Enzyme Research
Laboratories, Inc., was also used.
The inhibitory effects of the
peptides on integrin-ligand interactions were evaluated by using a
competitive enzyme-linked immunosorbent assay (ELISA) (17). Briefly,
for vitronectin-v
3 ELISA, 96-well microtiter plates were coated
overnight at room temperature with 10 µg/ml human vitronectin (Iwaki
Glass Co., Ltd., Tokyo), and the plates were washed with TBS containing
0.05% Tween 20.
v
3 solution was added to each well, and then the
peptide solution of various concentrations was added to each well. The plates were incubated at room temperature for 2 h and washed with TBS containing 0.05% Tween 20. Bound receptor was detected by using
anti-human
v rabbit polyclonal antibody (Chemicon Int., Inc.) and
affinity-purified goat anti-rabbit IgG conjugated to horseradish
peroxidase.
Fibrinogen-IIb
3 ELISA was performed using a similar protocol
except that microtiter plates were coated with 10 µg/ml human fibrinogen (Sigma) and anti-human
II mouse
monoclonal antibody was used.
To verify the specificities of ligand-integrin interactions, anti-v
monoclonal antibody (23C6; Serotec Ltd., UK) and anti-
IIb
3 monoclonal antibody (Becton Dickinson) were used.
The binding of the FITC-labeled RGD peptide to activated platelets was studied by flow cytometry. The activated washed platelets were prepared according to the methods of Frojmovic et al. (24) with some modifications. Human PRP was treated with 0.5 µM prostaglandin E1 (Sigma) at room temperature for 10 min. The pH of the PRP was then adjusted to 6.5 with 1 M citrate solution, and the PRP was centrifugated at 1200 × g for 10 min. The pellet was suspended in Ca2+-free Tyrode's solution containing 0.35% bovine serum albumin and 0.2 units/ml apyrase (grade III; Sigma), and the suspension was centrifugated at 1200 × g for 10 min again. The pellet was resuspended in Tyrode's solution containing 0.35% bovine serum albumin and 0.38% trisodium citrate (termed PRP (1:10)). 100 µM ADP was added to this platelet suspension and incubated at room temperature for 5 min. The platelets were then treated with 0.1 mM FITC-labeled peptide (FITC-WSRGDW) for 25 min at room temperature in the presence or absence of inhibitory peptides at various concentrations. The relative intensity of fluorescence was analyzed in a flow cytometer (Cyto ACE-300, Jasco, Tokyo, Japan).
Cell Adhesion AssayThe adhesion assay was performed in 24-well tissue culture plates. Each well was precoated with 30 µg/ml fibronectin (purified from human plasma, Iwaki Glass Co., Ltd., Tokyo) or 3 µg/ml vitronectin (purified from human plasma, Iwaki Glass Co., Ltd., Tokyo) at 4 °C for 24 h. On the day of the experiment, each well was blocked with 2% bovine serum albumin for 1 h at room temperature.
ECV 304 cells (obtained from Health Science Research Resource Bank,
Osaka Japan), a cell line cloned from human umbilical vein endothelial
cells (25), were cultured in medium 199 supplemented with 10% fetal
calf serum and 50 IU/ml penicillin and 50 µg/ml streptomycin at
37 °C. They were harvested using 0.05% trypsin and 0.5 mM EDTA and washed twice with divalent cation-free
phosphate-buffered saline. Finally, they were resuspended in serum-free
Eagle's minimum essential medium (5 × 105 cells/ml). The cells
were preincubated with appropriate inhibitors for 30 min at 37 °C,
and then 200 µl of this cell suspension then was added to each well
of the 24-well plates precoated with fibronectin or vitronectin, as
mentioned above, and were incubated at 37 °C in a CO2
incubator. After 30 min, each well was washed with serum-free Eagle's
minimum essential medium to remove unattached cells, and 100 µl of 65 µg/ml WST-1 (2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium monosodium salt) and 100 µl of 7 µg/ml
1-methoxy-5-methylphenazinium methylsulfate were added to each well.
After 1 h, the number of adhered cells was quantified by measuring
the absorbance at 450 nm (26). Anti-human v
3 monoclonal antibody,
23C6 (Serotec, Oxford, UK), was used to examine the contribution of
v
3 integrin (27). EILDVPST peptide (Funakoshi, Tokyo, Japan),
originating from the CS-1 site of fibronectin (28), was used to examine the contribution of the CS-1 site to the cell adhesion to
fibronectin.
Male Hartley guinea pigs weighing 230-320 g were anesthetized with pentobarbital (35 mg/kg, intraperitoneal), and an intratracheal cannula was inserted through the incision. To maintain a moderate anesthesia during the experiment, additional pentobarbital administration (intraperitoneal) was done if necessary. The polyethylene tubes filled with saline containing 100 units/ml heparin were inserted into the left jugular vein and right carotid artery, and these two tubes were connected with a silicon tube filled with the same solution. The whole length of this shunt was 6 cm. Each agent was infused via a cannula in a right jugular vein at the rate of 2 ml/h for 70 min. 10 min after the onset of the infusion, the silicon tube was exchanged with a polyethylene tube in which a silk thread was inserted. At every 20 min, the silk thread was exchanged, and the wet weight of the thread was measured. After the termination of the infusion, the thrombus formation was examined for another 60 min. The net thrombus weight was calculated from the thread weights before and after the experiments, and the total thrombus weight during 60 min was the sum of three net thrombus weights during each 20 min.
Peptide SynthesisAll peptides were synthesized by solid
phase methods using 9-fluorenylmethoxycarbonyl amino acids on an
automatic peptide synthesizer (model 431A, Applied Biosystems, Inc.).
The final step of the solid phase synthesis involved the acylation of
the N termini with different types of carboxylic acids in the presence of N,N-dimethyl formamide and
1-hydroxybenzotriazole in
1-(3,4
-dichlorophenyl)-2-isopropyl-aminoethanol. A
trimethylsilylbromide-thioanisol system was used for the final deprotection. The crude peptides were purified by reverse-phase HPLC
using a C18 preparative column. All peptides were characterized by fast
atom bombardment mass spectrometry or electrospray ionization mass
spectrometry, an amino acid analysis, and an analytical high pressure
liquid chromatography. The purity of the peptides was >98%. Labeling
of a peptide with FITC was performed using FITC isomer I
(Sigma). Free FITC was eliminated by dialysis.
The conformational analysis was conducted on Macro Model software, version 3.5d (29). Fifty thousand conformers were generated by the Monte Carlo method (30) and minimized with a modified Amber* Force field (31) in vacuum to obtain 2200 conformers. The distance between the guanidino group of the Arg residue and the carboxy group of the Asp residue (d1) and the distance between the imino group of the Pro residue and the carboxy group of the Asp residue (d2) were calculated. Conformers with shorter d1 or d2 (<10 Å) were eliminated, based on Kessler's active conformation (32). The electrostatic cut-off distance was set to 5 Å, and the force field was modified in order to not evaluate hydrogen bonds.
StatisticsThe statistical significance of the difference between two groups was determined by the Student's t test. The statistical significance of the difference between more than two groups was assessed by a one-way analysis of variance followed by Dunnet's test. Differences were considered to be statistically different when p was <0.05. Dose-inhibition curves were obtained by fitting the data to a logistic curve by the least squares method, and the IC50 values were calculated from the fitted curves.
In the previous study
(33), we found that an RGD-containing hexapeptide, PSRGDW, was a very
potent inhibitor of human platelet aggregation (IC50 = 0.79 µM). The Pro residue at the N-terminal position of this
peptide was essential for its high potency, because the substitution of
other amino acids for proline or removal of this Pro residue resulted
in a decrease in the anti-platelet activity (e.g. SRGDW was
50 times less potent than PSRGDW). These results suggest that the
N-terminal Pro residue is responsible for another interaction between
this peptide and the IIb
3 receptor besides the guanidino group
and the carboxyl group of the RGD site. However, this peptide was not
specific for platelet
IIb
3 integrin, although it was very potent.
To develop a potent antagonist of platelet
IIb
3 with high
specificity, we synthesized a series of peptides without the Arg
residue. The anti-platelet aggregation activity of these peptides is
shown in Table I. A replacement of the Arg residue at
the third position with norvaline, which means the removal of the
guanidino group of the Arg residue, resulted in a slight increase in
the anti-platelet activity, suggesting that the Arg residue of the RGD
site is not necessary for the anti-platelet activity of this peptide.
L-Proline, L-hydroxyproline, and
L-dehydroproline at the third position also reproduced the
anti-platelet activity of high potency. We also synthesized N-acetyl
PSPGDW, in which the imino group of the N-terminal Pro residue was
blocked, to investigate the role of this group and found that the
acetylation of the imino group resulted in the complete loss of the
activity, suggesting that this free imino group is essential to the
anti-platelet activity of this peptide. Among the peptides we
synthesized here, PS-Hyp-GDW (Hyp representing
4-hydroxy-L-proline) (NSL-9511) was the most potent
inhibitor of the platelet aggregation, and this peptide was used for
further experiments.
|
The anti-platelet activity of NSL-9511 differs apparently among species (Table II). Although NSL-9511 was a potent inhibitor of platelet aggregation in guinea pig PRP, it was much less active in mouse and rat PRP.
|
In our preliminary studies using flow cytometry, we have
found that NSL-9511 inhibited the binding of fibrinogen to
ADP-activated platelets, suggesting that the anti-platelet activity of
NSL-9511 results from the blockage of fibrinogen binding to its
receptor on the platelet surface. To clarify the mechanism underlying
the anti-platelet activity further, we performed a competitive ELISA in
which the binding of purified human platelet fibrinogen receptor, IIb
3, to immobilized human fibrinogen was detected by specific antibody.
v
3 ELISA was also performed using
v
3 purified
from human placenta to examine the binding specificity of NSL-9511. As
Fig. 1a shows, 100 µM NSL-9511
almost completely inhibited the binding of
IIb
3 to fibrinogen,
whereas it did not inhibit the binding of
v
3 to vitronectin. An
RGD-containing peptide, GRGDS, at the same concentration did inhibit
both the binding of fibrinogen to
IIb
3 and that of vitronectin to
v
3 completely. These results suggest that NSL-9511 is highly
specific for platelet
IIb
3. Anti-
v monoclonal antibody and
anti-
IIb
3, both of which have been reported as binding-blocking
antibody, completely inhibited the binding of vitronectin to
v
3
and fibrinogen to
IIb
3, respectively. As Fig. 1b
shows, NSL-9511 inhibited the binding of fibrinogen to
IIb
3 in a
dose-dependent manner with an IC50 value of 70 nM. On the other hand, this compound hardly inhibited the
binding of vitronectin to
v
3 and the percentage of inhibition was
only 23% at the highest dose (0.6 mM).
Effects on Binding of RGD Peptide to Activated Platelets
Although NSL-9511 is a potent antagonist of IIb
3,
it does not possess the RGD sequence. To determine whether NSL-9511 and the RGD peptide are mutually exclusive upon binding to
IIb
3, we
synthesized FITC-labeled RGD peptide and examined the effects of
NSL-9511 on the binding of this RGD peptide to activated platelets. As
Fig. 2 shows, activation of washed platelets with ADP
increased the mean fluorescence when the platelets were incubated with
100 µM FITC-WSRGDW. This increment in mean fluorescence
was inhibited by the addition of excess unlabeled RGD peptide (1 mM WSRGDW) (data not shown), suggesting that it resulted
from the specific binding of FITC-labeled WSRGDW to platelet integrins
in an activation-dependent manner. An addition of 10 µM NSL-9511 completely inhibited the increase in the mean
fluorescence. These results suggest that NSL-9511 and the RGD peptide
share the same binding site on
IIb
3 or that their binding sites
are very close.
Effects on Cell Adhesion to Vitronectin and Fibronectin
The
binding specificity of NSL-9511 was examined in a cell adhesion assay,
in which ECV304 cells, originating from human umbilical vein
endothelial cells, adhered to immobilized vitronectin and fibronectin
via integrin receptors. As shown in Fig. 3a,
the adhesion of ECV 304 cells to immobilized human fibronectin was
inhibited by the GRGDS peptide, and almost complete inhibition was
observed at 1000 µM. The EILDVPST peptide, which is the
sequence within the CS-1 site of fibronectin and is considered to be
another sequence responsible for the adhesion of various cell types to
fibronectin (28), did not show such an inhibitory activity up to 100 µM under this assay condition. These results suggest that
the adhesion of ECV 304 cells to fibronectin will be mainly mediated by
the 5
1 (fibronectin receptor)-RGD interactions. On the other
hand, the adhesion of ECV 304 cells to immobilized vitronectin was
inhibited by GRGDS by about 50% even at 1000 µM (Fig.
3b). An excess amount of 23C6, the monoclonal antibody to
v
3 that has been reported to inhibit cell adhesion via
v
3
integrin (27), also inhibited the ECV304 adhesion for vitronectin by
about 50%. In this adhesion assay, NSL-9511 did not inhibit the
adhesion of ECV 304 cells to immobilized fibronectin and vitronectin up
to 1000 µM (Fig. 3, a and b),
suggesting that NSL-9511 does not have inhibitory actions on integrins
such as
v
3 and
5
1 and that it is highly specific to
platelet
IIb
3.
Guinea Pig Arteriovenous Shunt Model
To examine the
antithrombotic activity of NSL-9511 in vivo, a guinea pig
arteriovenous shunt model was used. In this model, a shunt circuit is
constructed between a carotid artery and a jugular vein with a silicon
tube in which a silk thread is inserted, and the circulation of blood
in this shunt results in a platelet-rich thrombus formation around the
silk thread. As shown in Fig. 4a, the total
thrombus weight over 60 min was 14.8 ± 1.2 mg in the control
group where saline was infused via a jugular vein. Infusion of heparin
(100 units/kg/h), an anticoagulant, inhibited the thrombus formation by
about 50%. NSL-9511, administered by infusion, inhibited the thrombus
formation in a dose-dependent manner, and at a dose of 10 mg/kg/hr, no thrombus was formed at all during the infusion. After the
termination of the infusion, the thrombus formation was examined for
another 60 min (Fig. 4b). Although the antithrombotic effect
of heparin continued for at least 60 min after the termination of its
infusion, the effect of NSL-9511 disappeared quickly, and the thrombus
weight formed during this 60 min was not significantly different from
that of the control group even after the administration of 10 mg/kg/h
(13.9 ± 1.7 mg in the control group versus 13.1 ± 0.65 mg in the 10 mg/kg/h group). These results suggest that NSL-9511, infused via the jugular vein, inhibited the platelet-rich thrombus formation in vivo, but the duration of its
antithrombotic action is very short. RGDS did not inhibit thrombus
formation at a dose of 100 mg/kg/h in this model.
We have been developing specific antagonists of integrin receptors
such as IIb
3,
v
3, or
5
1 using small linear peptides as the basic structure, because they are, in general, easy to synthesize by standard solid phase methods, and we can obtain much
information about the structure-activity relationship by changing the
constituent amino acids. In this study, we have found a series of
potent and highly specific antagonists of platelet
IIb
3 integrin
that are linear hexapeptides without the RGD sequence. Although several
peptide
IIb
3 antagonists with high specificity for this integrin
have already been reported, such as fibrinogen
-chain C-terminal
decapeptide (19, 20), a tick saliva protein (21), barbourin (22), and
synthetic cyclic peptides (16-18), this is the first report of small
linear ones using L-form natural amino acids. Scarborough et
al. (22) have reported that barbourin, a member of the disintegrin
family of snake venom proteins, inhibits the binding of fibrinogen to
IIb
3 but does not block vitronectin binding to
v
3.
Barbourin has a KGD sequence rather than RGD; therefore, the Lys
residue is thought to be critical for the binding specificity. These
results suggest that for potent antagonism toward integrins that
recognize a RGD sequence such as
IIb
3,
v
3, or
5
1, two
interacting sites, an acidic moiety and a basic moiety, are necessary,
and the substitution for a guanidino group of the Arg residue by the
other basic groups can produce the binding specificity. Our results
described here also demonstrate that the guanidino group of the Arg
residue within the RGD sequence is not unique to the antagonistic
activity toward
IIb
3, that the imino group of proline may
displace the role of the guanidino group, and that the introduction of
the Pro residue to the N terminus of the peptide results in more potent
activity and binding specificity.
Now we are studying the close structure-activity relationships around this peptide. Because the N-terminal Pro residue and the Asp residue at the fifth position are essential to the activity, the basic structure of the peptides can be summarized to have the following motif, Pro-X1-X2-X3-Asp-X4. The results of the preliminary experiments suggest that 1) a small amino acid such as Ser, Ala, or Gly is preferable at the X1 position, 2) X2 may be any amino acid, 3) X3 must be a small amino acid with a residue such as Gly, or a cyclic amino acid such as Pro, and 4) X4 prefers an amino acid with an aromatic side chain. We are going to discuss the structure-activity relationship of this peptide in detail elsewhere.
One of the motif peptides, NSL-9511, inhibited the binding of
fibrinogen to purified human platelet IIb
3 in a
dose-dependent manner with an IC50 of 70 nM (Fig. 1). Because NSL-9511 also inhibited in
vitro platelet aggregation at submicromolar concentrations (IC50 = 390 nM), the anti-platelet activity of
NSL-9511 can be explained by the blockage of the fibrinogen binding to
IIb
3. Further, in this binding assay, the concentration of
fibrinogen in the incubation medium did not greatly affect the
IC50 value of NSL-9511 (data not shown). These results
suggest that NSL-9511 inhibits fibrinogen binding, not by interacting
with fibrinogen molecules but by binding to
IIb
3 integrin as an
RGD peptide does. Therefore, we concluded that NSL-9511 is an
IIb
3 antagonist.
Because NSL-9511 does not have an RGD sequence and is a much more
potent antagonist of IIb
3, we then examined whether the binding
mode of NSL-9511 and an RGD peptide is the same. The results of the
experiment showing that NSL-9511 dose-dependently inhibited the binding of an FITC-labeled RGD peptide to activated platelets (Fig.
2) suggest that NSL-9511 and the RGD peptide share the same binding
site on
IIb
3 or that the binding site of NSL-9511 lies so close
to the RGD-binding site that NSL-9511 and an RGD peptide are mutually
exclusive upon binding to
IIb
3.
These findings, that NSL-9511 is an IIb
3 antagonist and possibly
has the same binding mode as an RGD peptide and that the N-acetylation of the imino group of the Pro residue results
in the complete loss of the anti-platelet activity, lead us to the idea
that the imino group of the Pro residue at the N terminus replaces in a
complementary manner the function of the guanidino group of the Arg
residue within an RGD sequence as a basic moiety essential to the
activity. This idea is supported by the conformational analysis of a
PSRGDW peptide, which has both the imino group of the Pro residue and
the guanidino group of the Arg residue. The spatial distance between
the carboxyl group of the Asp residue and the guanidino group of the
Arg residue (d1) of stable conformers and the spatial distance between
the carboxy group and the imino group of the Pro residue (d2) of stable
conformers were calculated (Fig. 5). The largely
overlapped area of d1 and d2 distribution suggests that the relative
arrangement between the Pro residue and the Asp residue is similar to
that of the Arg residue, indicating that the imino group of the Pro
residue can occupy the same conformational space as the guanidino group
of the Arg residue.
To examine the binding specificities of the motif peptide, we performed
the binding assay using purified human placental v
3 and
vitronectin and the cell adhesion assay, in which ECV 304 adhered to
immobilized fibronectin or vitronectin. In the
v
3-vitronectin binding assay, both an anti-
v
3 monoclonal antibody and 100 µM GRGDS completely inhibited the binding, suggesting
that the binding was mediated by
v
3 and the RGD sequence of
vitronectin. The inhibition of the binding of vitronectin to
v
3
by NSL-9511 was very weak at its highest concentration (Fig.
1b), suggesting that NSL-9511 is highly specific for
IIb
3. The results of the cell adhesion assay (Fig. 3) also show
that NSL-9511 did not inhibit cell adhesion via other integrins such as
v
3 or
5
1. ECV304 is a cell line established from human
umbilical vein endothelial cells, and we have confirmed the expression
of
v
3 and
5
1 on ECV304 using monoclonal antibodies by flow
cytometry (data not shown). Adhesion of ECV304 to vitronectin is
thought to be largely mediated by
v
3 because the monoclonal
antibody to the
v
3 complex inhibited the adhesion. However, we
could not show the dependence of
5
1 on the adhesion of ECV304 to
fibronectin, because an anti-
5
1 antibody that has been reported
to block the binding of fibronectin to
5
1 is not available. These
results suggest that NSL-9511 is an antagonist highly specific for
platelet
IIb
3 but not for other integrins such as
v
3 or
5
1. There are several possible explanations for this high
specificity of NSL-9511 for
IIb
3. First, it is possible that
integrins other than
IIb
3 prefer a guanidino group as a basic
interaction site upon binding with the ligand, while both a guanidino
group and secondary amines such as an imino group of the Pro residue
are preferable as a basic moiety upon ligand binding to
IIb
3. It
is also possible that the active spatial structure for
v
3 or
5
1 integrin is quite different from that for
IIb
3 and that
the NSL-9511 cannot conform to this structure, while the RGD peptide
can.
NSL-9511 also exhibited the anti-platelet activity in vivo,
when infused intravenously (Fig. 4). Infusion of NSL-9511 at a rate of
10 mg/kg/h almost completely inhibited the platelet-rich thrombus
formation around the thread, and this antithrombotic activity quickly
wore off after the termination of the infusion. Peptides, including
NSL-9511 are rapidly degraded in guinea pig plasma at 37 °C by the
action of aminopeptidases (data not shown); therefore, this quick
disappearance of anti-thrombotic activity would be due to the decrease
in plasma concentration of NSL-9511 caused by the enzymatic
degradation. The same is true for an RGDS peptide, which did not show
any antithrombotic activity in this shunt model. Substantial increases
in bleeding time have been an undesirable characteristic exhibited by
many IIb
3 antagonists (34) including an anti-
IIb
3
monoclonal antibody, 7E3 (35, 36), and this is one of the limitation
factors in the use of an
IIb
3 antagonist as an antithrombotic
agent. Therefore, NSL-9511, which exhibits potent antithrombotic
activity in vivo during the infusion, although this activity
wears off quickly and the prolongation of bleeding time is not as
serious, has a potential for clinical uses, especially as an
antithrombotic drug at an acute phase.
Another interest of ours is the possibility that these motif peptides can be produced endogenously. Although many of the human proteins include the sequences homologous to this motif sequence, they do not lie at the N terminus, and the imino group of the Pro residues is not free. To exert anti-platelet activity, these proteins need to be processed so as to expose the Pro residue at a new N terminus. Aminopeptidase P, which has been reported to exist in human cells such as platelets (37), and some virus proteases such as HIV-1 protease (38) are candidates for catalyzing such a reaction.
We are grateful to Dr. Seiichi Ohkuma
(Emeritus Professor of Tokyo Collage of Pharmacy) and Dr. Setsuko
Yamazaki (Japanese Red Cross Tokyo Western Blood Center) for the
generous gift of outdated platelets and to Dr. Masaaki Ohkuma
(Institute of Life Science, Kurume University) for the generous gift of
platelet IIb
3-rich membrane fraction. We also thank Emiko Yasuda
and Mako Yano for technical assistance in peptide syntheses.
HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
All ASBMB Journals | Molecular and Cellular Proteomics |
Journal of Lipid Research | Biochemistry and Molecular Biology Education |