- Email: [email protected]
Inhibition of human platelet aggregation by GR91669, a prototype fibrinogen receptor antagonist
Thrombosis Research, Vol. 75, No. 3, pp. 269-284.1994 Copyright 8 1994 Elsevier Science Ltd Printed in the USA. All rights reserved GO49-3848/94 $6.00 + .OO
INHIBITION OF HUMAN PLATELET AGGREGATION BY GR91669, A PROTOTYPE FIBRINOGEN RECEPTOR ANTAGONIST Martyn R.Foster, Edward J.Hornby, Susan Brown, Michael Harm’, John Kitchinl, Nicholas Pike, and Peter Ward’. Departments of Cardiovascular and Respiratory Pharmacology and ‘Medicinal Chemistry,Glaxo Group Research Limited, Ware, Herts.
(Received 23 February 1994 by Editor M.E Scully; revised/accepted 5 May 1994)
Abstract
In order to produce more potent and specific fibrinogen receptor (GpIIb/IIIa) antagonists, the Arg-Gly of a chemical series based upon Arg-Gly-Asp was replaced by alkyl chains of varying lengths. The most potent in this series, GR91669, inhibited aggregation of human gel-filtered platelets (GFP) in vitro induced by ADP or the thromboxane A2 mimetic, U46619, with IC50 values of 200nM and 500nM respectively and was selected for further studies. Its inhibitory effects on GFP were reversed by addition of excess fibrinogen. The compound also inhibited ADP- or U466 19- induced latelet aggregation in human whole blood (IC50 values of 700nM in both cases). 12PI-Fibrinogen binding to ADP-stimulated platelets was inhibited by GR91669 with an IC50 ( 65nM) similar to that against platelet aggre ation. GR91669 (1mM) did not inhibit U46619-induced platelet shape change or 1PC-5HT secretion from platelets stimulated by collagen, U46619 or thrombin. Therefore GR91669 inhibits aggregation but has no significant effect on stimulus-response events, a profile consistent with fibrinogen receptor blockade. In addition, GR91669 (ImM), unlike echistatin or Gly-Arg-Gly-Asp-Ser, did not disrupt vitronectin receptor-dependent attachment of cultured HUVECS in vitro and similarly did not inhibit Mac-l dependent adhesion of human granulocytes. Thus, of the integrins tested, GR91669 appears to be specific for GpIIb/IIIa. Following intravenous administration to marmosets of 1 or lOmg/kg GR91669, ADP (lObM)-induced platelet aggregation ex vivo was abolished for 15 and 60 minutes respectively. Greater than 50% inhibition was maintained for 30 minutes and 2 hours respectively. GR91669, therefore appears to be a potent, specific fibrinogen receptor antagonist in vitro and which is also active in vivo.
Key words: GR91669, GR83895, echistatin, human platelets, Glycoprotein IIb/IIIa, Fibrinogen receptor antagonists, Integrins.
269
270
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
Platelets play a central role in thrombosis. After damage to the endothelial cell lining of blood vessels, proteins are exposed providing a surface on which platelets may adhere. Subsequently, adhered platelets become activated and through secretion of aggregatory agonists recruit further platelets to the injury site. During such activation, and as a result of exposure to aggregatory agonists, glycoprotein (Gp) IIb/IIIa, the fibrinogen receptor, changes its conformation and binds fibrinogen (1). Such binding results in the crosslinking of platelets by fibrinogen leading to aggregation and thrombus formation. This may result in obstructed blood flow with potentially fatal consequences.
Current anti-thrombotic therapies prevent the action of a single aggregatory agonist, but do not attenuate the effects of other agonists. For example, aspirin prevents the formation of thromboxane A2 but does not prevent aggregation induced by other agonists such as adenosine diphosphate (ADP) or thrombin (2). However, the binding of fibrinogen to GpIIb/IIIa represents a common event during platelet aggregation, regardless of the initiating stimulus. Thus, antagonism of platelet fibrinogen receptors inhibits platelet aggregation to a broad spectrum of aggregatory agonists, and represents a pharmacological approach more effective than current therapies to the potential treatment of thrombosis.
The amino acid sequence, Arg-Gly-Asp (RGD) is present in the Aa chain of fibrinogen and binds to GpIIb/IIIa (3). It is predicted that RGD exists in a p-loop conformation within fibrinogen (4). We have synthesised and previously reported results obtained with cyclicpeptides based upon RGD. GR83895, a synthetic cyclic peptide, causes inhibition of fibrinogen binding and platelet aggregation with improved potency compared with homologous linear peptides (5). GpIIb/IIIa is a member of the integrin super-family of adhesion receptors and the GpIIIa (p3) subunit is common to the fibrinogen, vitronectin and fibronectin receptors. Other fibrinogen receptor antagonists such as GRGDS and echistatin appear to have additional inhibitory effects on vitronectin receptors, whilst GR83895 is lOO-fold more specific for fibrinogen receptors (5). However, GR83895 lacks the potency of other fibrinogen receptor antagonists such as echistatin, and therefore we have synthesised a series of linear molecules in an attempt to obtain more potent fibrinogen receptor antagonists with similar specificity of action.
This study describes the in vitro structure activity relationships of a series of novel linear peptidic compounds based upon RGD, and the in vitro pharmacological profile and specificity of action of the most potent compound, GR91669. Previous reports have suggested that platelets from several commonly used laboratory species, including rat and rabbit, are not inhibited by fibrinogen receptor antagonists (6). After preliminary in vitro investigations, marmoset platelets were found to be similar to human platelets in sensitivity to inhibition by fibrinogen receptor antagonists, and therefore, we also describe the potency and duration of action of GR91669 after intravenous administration in this species.
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
271
MATERIALS AND METHODS
Materials
GR91669 (8-[(aminoiminomethyl)thio]-L-aspartyl-L-phenylalanine), GR83895 (5 ), compounds I-VIII (as trifluoroacetate salts) and prostacyclin were synthesised in the Chemistry Division, Glaxo Group Research. GR91669 was prepared as a 1mM stock solution in 5mM Hepes in 0.9% w/v sodium chloride solution (saline) (pH 8.4) and diluted with 5mM Hepes in saline (pH7.4) for use. Compounds I-VIII were dissolved in dimethylsulphoxide (DMSO) and diluted in 5mM Hepes buffered saline (pH 7.4) to give a stock concentration of 1mM in 10% v/v DMSO. GR83895, Gly-Arg-Gly-Asp-Ser (GRGDS; Bachem AG, Bubendorf, West Germany) and echistatin (Bachem (UK) Ltd) were prepared as 1mM stock solutions in 5mM Hepes buffered saline (pH 7.4). Prostacyclin was prepared as a 1mM stock solution in 50mM Tris buffer (pH 9.0), stored at -20°C and was diluted to working concentrations in 50mM Tris buffer (4’ C, pH 8.0) on the day of use. Aspirin (Sigma Chemical Co.) was prepared as a lOmg/ml solution in 1OOmM Tris buffer (pH 8.5). Adenosine diphosphate (ADP; Sigma Chemical Co.) was dissolved in 50mM Tris buffer (pH 6.5) as a 1mM stock solution. U46619, a chemically stable TxA2 mimetic (Cayman Chemicals) was prepared as a 1mM stock solution in 1% w/v NaHC03 in saline. Equine collagen suspension was from Hormon Chemie, Munich, West Germany. Human cr-thrombin (Sigma Chemical Co.) was prepared as a 10 IU/ml solution in saline. Ethylene glycolbisaminoethylether tetra-acetic acid (EGTA) was prepared as a 400mM stock solution in 20mM Hepes buffered saline (pH7.4) and adjusted to pH7.4 by addition of 5N NaOH. yF5rnan fibrinogen was from Kabt.’ 14C-5-hydroxytryptamine (14C-5HT; SOmCi/mmol) and Iodine were from Amersham International Plc. Penicillin, gentamycin, streptomycin, Hank’s medium, medium 199 (M199), foetal calf serum and glutamine were from ICN-Flow, High Wycombe, UK. Collagenase was from Boehringer Mannheim, Lewes, UK. Rose Bengal stain was from Koch-Light Ltd, Haverhill, UK. Monoclonal antibodies to CD1 1 (anti-Macl) and GpIIIa were from Dakopatts, Denmark. Hank’s buffered salt solutions (HBSS) and sterile Hepes, bovine serum albumin and sodium bicarbonate solutions were from Gibco. All other reagents used during the preparation of granulocytes were from Sigma Chemical Company. All other reagents were from BDH Ltd., Poole and were of analar grade.
Platelet preparation
Blood from human volunteers who had taken no medication for 10 days was collected by venepuncture into anticoagulant (9 vol. blood: 1 vol. 0.13M trisodium citrate solution). Plateletrich plasma (PRP) and gel-filtered platelets (GFP) were prepared as described previously (7). Fibrinogen was added to GFP to a final concentration of 5OO&ml (5) except in experiments where competition studies were performed.
Platelet aggregation
Aggregation of GFP was induced by addition of ADP (10pM) and studied using a 6 channel, optical aggregometer (Glaxo Bioengineering Unit) as described previously (5). Increasing concentrations of potential inhibitors of aggregation were incubated with GFP for 5 minutes at 37 OC, prior to addition of ADP. Aggregation was expressed as the percentage change in light
272
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
transmittance compared with a vehicle-treated, aggregated sample. IC50 values were calculated as the concentration of compound which caused 50% inhibition of aggregation.
Platelet aggregation in human titrated, whole blood was quantified by counting the fall in single platelet count after addition of U46619 (3pM) or ADP (301rM), according to a method described previously (5). IC50 values were calculated for each of the compounds tested, as the concentration of inhibitor which caused a 50% inhibition of the fall in platelet count measured.
12jI-Fibninogen binding studies. GRGDS, GR91669 and echistatin were tested as inhibitors of 1251-fibrinogen (2pCi, 5OO&ml final cont.) binding to ADP (20pM -stimulated GFP, according to the method described previously (5). Bound and unbound 1251-fibrinogen were separated by centrifugation through dioctyl phthalate / dibutyl phthalate (6:4 v/v) mixtures and the radioactivity in the platelet pellets estimated by scintillation counting.
Plateletshape change Platelet shape change was measured optically in a 6-channel aggregometer as the decrease in light transmittance through a sample of aspirin-treated PRP. By amplifying the output from the aggregometer it was possible to record shape change. Platelet aggregation was prevented by addition of EGTA (4mM) lmin prior to the induction of platelet shape change with U46619 (0.3~M). Compounds were incubated with aliquots of stirred (90Orpm) PRP (37’C) for 5 mins before agonist addition. The peak decrease in light transmittance was measured and expressed as a percentage of that obtained in a vehicle-treated sample.
14C-5HT secretion 14C-5HT secretion was measured in PRP or GFP prepared from blood, incubated (37°C 30 min) with 14C-5HT (lccCi/20ml), according to the methods described previously (7). GR91669 or prostacyclin concentration-inhibition curves were constructed in PRP against platelet aggregation and concurrent l4 C-5HT secretion induced by U46619 (lOpM), collagen (2rrglml) or ADP (10pM). Similarly, concentration-inhibition curves for GR91669, prostacyclin or echistatin were constructed in GFP against thrombin (O.lIU/ml)-induced platelet aggregation and 14C-5HT secretion. In order to avoid thrombin-induced clotting, fibrinogen was omitted from the suspension medium of GFP. 14C-5HT secretion was expressed as a percentage of the total 14C5HT present in an equal volume of PRP.
Competitionstudies Six cuvettes containing GFP, prepared without addition of fibrinogen, were incubated without stirring for 5 min at 37’ C, with a single concentration of inhibitor or vehicle. A different concentration of fibrinogen over the range l-2000~g/ml was then added to each cuvette and stirring (900 rpm) initiated. After 1 min, ADP (lOtiM) was added and the maximum aggregation occurring within 10 min was measured. A series of such aggregation curves was generated in a
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
randomised order, with up to three different concentrations way in each experiment.
273
of inhibitor examined in a similar
Detachment of human umbilical vein endothelial cells (HUVECs) Inhibitors were tested for their ability to detach HUVECs according to the method described previously (5). Compounds, prepared in culture medium, were incubated (2 hours, 37°C) with HUVEC cultures grown to confluence in 24 well plates. After washing the wells with fresh culture medium, the density of the remaining cells was estimated using a staining technique (5) and the results expressed as percentage detachment relative to wells containing HUVECs, which had not been exposed to compounds. In separate experiments, cultured HUVECs grown to confluence, were treated with GRGDS (1mM) or echistatin (1 BM) for 2 hours (37’C). Detached cells were harvested and washed in culture medium, before plating onto fresh culture plates and incubated (37°C under 5% v/v C02) until confluent.
Granulocyte adhesion Human whole blood was collected into anticoagulant (130mM trisodium citrate, 1mM sodium dihydrogen orthophosphate, 1lmM glucose; 6.7 vol. blood : 1 vol. anticoagulant). Red blood cells were removed by sedimentation through methyl cellulose (40m1, 13% w/v) containing sodium diatrizoate (0.6% w/v). The plasma fraction containing the granulocytes was retained and subsequently the granulocytes purified by centrifugation (4OOxg, 20 mins) on Ficoll-Paque. Contaminating red cells were removed by lysing with 0.2% w/v sodium chloride. After addition of 1.6% w/v sodium chloride to restore isotonic balance, the granulocytes were washed by centrifugation and resuspension in calcium free-Hank’s buffered salts solution. The granulocyte concentration was determined using a Coulter JT haematology analyser and adjusted to give a final concentration of 1~10~ cells /ml.
Purified human fibrinogen (300&ml) was dispensed into 24 well plates and incubated (17-20 hrs, 4°C) prior to use. The fibrinogen solution was replaced with HBSS containing 0.1% bovine serum albumin and incubated for lhr at 20°C. Each well was aspirated, test compound or vehicle (100~1) added and equilibrated (37°C) in a Dynatech shaker-incubator prior to the addition of granulocyte suspension (lml). After mixing for 10 mins, formyl-met-leu-phe (fMLP) (5nM) was added and incubated for 30mins (37°C) without shaking. Plates were agitated for lmin, each well aspirated, and lml of HBSS added and the procedure repeated. Adhered granulocytes were lysed by addition of cetyltrimethylammonium bromide (CTAB 0.3% w/v, 250~1; 30mins) and the lysate assayed for alkaline phosphatase, according to the method of Santini et al (8). Alkaline phosphatase values were expressed as a percentage of that obtained from a lml sample of cell suspension.
Marmoset whole bloodplatelet aggregation. Blood was collected by venepuncture from the femoral artery of conscious, restrained marmosets Callithrix jacchus (300-500gm) of either sex and anticoagulated as for human blood. Aliquots (150~1) of blood were stirred in an aggmgometer (lOOrpm, 37°C) and a sample (10~1) taken from each for control platelet counting using an Ultra-F10 100 platelet counter (Becton Dickinson).
274
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
After 5 minutes incubation (37’C) in the presence of GR91669 (O.l-lOtiM), echistatin (0.03IrrM), GRGDS (lo-1OOOpM) or vehicle, aggregation was induced by addition of a justmaximally effective concentration of ADP (30pM), and a further sample taken for platelet counting 1 minute after agonist addition. Prior to platelet counting, the (10~1) blood samples were added to 2ml of formal-saline (0.2% w/v), containing ethylene-diamine-tetraacetic acid (EDTA; 1mM). The percentage fall in platelet number was calculated relative to the control platelet count prior to addition of agonist. Inhibition of aggregation was expressed by comparing the percentage fall in platelet number, with that obtained in a maximally aggregated (30pM ADP) sample.
EX vivo whole blood platelet aggregation was performed in a similar manner. GR91669 (1 or lOmg/kg) or vehicle was administered to marmosets (300-5OOg) of either sex by intravenous injection, via a hind leg vein. Blood (0.6ml) was collected from a femoral vein, before, and at intervals up to 4 hours after administration of compound. Aliquots of blood (150~1) were dispensed into four cuvettes and ADP (l-30pM) concentration-effect curves of aggregation constructed as described above. RESULTS
Structure-activitystudies. We have previously shown that the potency of RGD-like molecules as inhibitors of platelet aggregation could be improved by cyclisation (5). In the search for more potent and specific fibrinogen receptor antagonists a series of linear peptides based upon RGDF were synthesised. When tested in GFP or whole blood, all the peptides caused complete inhibition of platelet aggregation induced by ADP (10pM) or U46619 (3pM). GRGDS was a relatively weak inhibitor, particularly when measurements were made in whole blood (Table 1). Removal of the N-terminal glycine and substitution of serine with phenylalanine gave RGDF which was 3-fold more potent as an inhibitor of ADP-induced platelet aggregation in GFP (IC50 8pM; Table 1). In an attempt to achieve further increases in potency, modifications were made to the linear RGDF sequence. The assumption that the interaction of the arginine with the fibrinogen receptor was through its alkyl-guanidine side chain was supported by the observation that activity was retained in compound I (ICso 5.1 bM), in which the N-terminal amino group of arginine was removed, but alkyl-guanidine retained (Figure 1). Further modification of compound I showed that the guanidine could be replaced by a simple primary amino group (compound II; IC50 14gM), with only a 2-3-fold loss in potency. In this series of more readily prepared compounds, we investigated the effect of different alkyl chain lengths (3-7 carbon atoms), thereby varying the distance between the terminal amino and carboxylic acid groups. As can be seen from Figure I (compounds II-VI), a chain length of 6 carbons produced optimal activity (compound V, IC50 1.8pM). Having optimised the alkyl chain length, the primary amino group in compound V was replaced with isothiourea, a group which more closely resembles a guanidine function, in order to regain the activity lost in compound II. However, to maintain the optimal length of compound V it was necessary to shorten the alkyl chain from 6 to 4 carbons, thus compensating for the additional length of isothiourea. These changes resulted in a 40-fold increase in potency (compound VII, IC50 0.3pM) over that of compound I. Having succesfully replaced the arginine of RGDF and enhanced potency we turned our attention to the possibility of replacing glycine.
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
275
Glycine often acts only as a peptide spacer and therefore we recognised that again it might be replaced by a simple non-peptidic alkyl chain spacer. Replacement of the glycine peptidic bond of compound I by an alkyl chain gave compound VIII, which had a 4-fold increase in potency (IC50 1.2~M). Finally, the guanidine group of compound VIII was again replaced by an isothiourea to give GR91669 (IC50 0.2~M), the most potent compound of this series. In comparison with the highly potent snake venom peptide, echistatin, GR91669 was 2-4 times less potent in GFP and 3-9 times less potent in whole blood (Table 1). In whole blood, the rank order of potency for the compounds tested was maintained, but with a tendency for the weaker compounds to decrease slightly in potency against a given aggregatory agonist, relative to GFP suspensions (Table 1).
X-(CH&CO-Gly-Asp-Phe Compound
X
IC50 (ctM)vs ADP
I
Guanidine
5.1
II
14.0
III
45.0
IV
64.0
V
1.8
VI
3.0
VII
0.3
Isothiourea
Y-(CH$,CO-Asp-Phe Compound
Y
n
IQ0 (PM)
VIII
Guanidine
7
1.2
GR91669
Isothiourea
7
0.2
FIG. 1 Structures of compounds I-VIII and GR91669 related to inhibitory potency.
Fibrinogen binding studies. Each of the inhibitors tested prevented the binding of fibrinogen to ADP (lObM)-stimulated platelets (Figure 2). Echistatin and GR91669 had IC50 values of 0.05~M and 0.06bM
276
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
respectively. GRGDS was significantly less active with an IC50 of approximately 2OrrM, although complete inhibition of binding could be achieved at concentrations of 30Or1M. 100 g
go-
2 ho_ ._ A7 : 408
20o.OOl .Ol .l 1 10 Concentration (FM)
100
1000
FIG. 2 The effect of GR91669 (O,O), GRGDS (QH) or echistatin (A,A) on the ADP(lOrM)induced binding of 1251-fibrinogento human platelets (solid symbols) and aggregation (open symbols) in GPP. Points are mean and S.E.M. of at least 3 independent observations. Where not shown error bars arc within the symbols.
TABLE I The effect of a range of compounds on human platelet aggregation induced in GPP or blood after addition of ADP or U46619. Compound
GFP IC50 (PM) ADP(lOuM)
U46619(3pM)
ADP(30pM)
U46619(3fiM)
(1:;6) NT
132 (79-222) NT
372 (206-687) NT
GRGDS RGDF
v VII VIII
(&I) 8
NT
(O.lY4.0, 0.3 (0.1-1.6) 1.3 (1.0-1.7)
NT
(O.Z.3) 0.05 (0.04-0.06)
(O.Z7) 0.3 (0.06- 1.6)
NT
GR91669 Echistatin
Whole blood IC50 (PM)
3.1 (1.4-7.1) (OE.3) 5.9 (3.5-9.9) 0.7 (0.6-1.0) 0.08 (0.05-O. 12)
(l&3) (3.7Y7.8) 12.3 (6.7-22.9) (o.:~:.o) 0.2 (0.2-0.3)
Values are Geometric means and 95% confidence limits on at least 3 separate determinations. NT-not tested.
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
(b)
277
(4
FIG. 3
Typical aggregometer traces illustrating (a) shape change induced by U46619, (b) effects of GR91669 (1mM) and (c) prostacyclin (0.1~M) on U46619 (0.3rcM)-induced shape change in the presence of EGTA (4mM). Shape change studies
None of the inhibitors of platelet aggregation caused shape change on addition to PRP. Additionally, shape change induced by U46619 (0.3ctM) was not inhibited by any of the peptide analogues even at concentrations up to 1mM. In contrast, prostacyclin was potent in inhibiting shape change, with an IC50 value of 0.04pM (Figure 3). 14C-SHT secretion
U46619 (lOlM), collagen (2&ml) or ADP (10uM) induced platelet aggregation and concurrent secretion of 14C-5HT from stirred PRP (Figure 4). U46619- and collagen-induced secretion was not inhibited by GR91669 (O.l-3.0pM; Figure 4b) whilst, in contrast, ADP-induced secretion was totally inhibited. In the same PRP samples, GR91669 (0.3-3.0 PM) caused 90-100% inhibition of the concommitantly measured platelet aggregation induced by each of the agonists (Figure 4a). Both GR91669 (0.01-I PM) and echistatin inhibited thrombin-induced aggregation, but not 14C-5HT secretion, in GFP, whereas prostacyclin (lOnM-1OOnM) inhibited aggregation and secretion to all agonists in parallel, as illustrated against thrombin (Figure 4c & 4d).
Competition studies.
When added to ADP (lOpM)-stimulated platelets, fibrinogen caused a concentration-related increase in platelet aggregation as measured by optical aggregometery. GR91669 and GR83895 caused concentration-related, parallel, rightward shifts of fibrinogen concentration-effect curves (Figure 5). Echistatin at 0.1 PM caused an apparently unsurmountable inhibition of aggregation. It is not known whether higher concentrarions of fibrinogen might have restored the aggregation response.
Specificity studies
GR91669 did not cause detachment of HUVECs at concentrations up to 1mM (data not shown). Echistatin and GRGDS were both able to cause complete detachment of the cells at
278
(a)
(b) 100
+ + *
U46619 (10pM) ADP (10~M) Collagen (2pg/ml)
1
.l
r
Concentration GR91669 (FM)
(4
w
-
-& +
20 0
Prostacyclin CR91669
.g H $
80
g 3 ::
40
60
20 I
J
t
0 1
.OOl
.Ol
10
1
.l
10
Concentration GR91669 (FM)
e
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
I
I
.l
1
L I
.OOl
.Ol
I
I
.l
1
Inhibitor Concentration (FM)
Inhibitor Concentration (PM)
FIG. 4 The effect of GR91669 on aggregation and concurrently measured 14C-5HT secretion induced by U46619 (lObM), ADP(lOkM), or collagen (2&ml) in PRP (a Jr b) and induced by human thrombin (0.1 W/ml) in GFP (c & d). The results are the mean and S.E.M of 4 experiments. Error bars omitted in figure 4c for clarity. concentrations of 0.1 crM and 100~M respectively, with IQ values of 0.01 BM and 48pM (data not shown). Approximately 40% of cells remained attached in the presence of GR83895 (100~M). Cells which had been displaced by echistatin or GRGDS were harvested, washed and grew to confluence within a normal time scale, suggesting that detachment was not due to intrinsic cytotoxicity in either peptide. Granulocyte adhesion Of the granulocytes added to each well, approximately 20-40% became adherent after 30 min of incubation. None of the inhibitors tested caused any reduction in fMLP-stimulated granulocyte adhesion (data not shown). A monoclonal antibody to Mac-l (l/1000-l/10 v/v) caused concentration related reductions in granulocyte adhesion, with complete abolition at titres of l/10 v/v (data not shown). A monoclonal antibody of the same immunoglobulin class as that to Mac1 had no inhibitory effect.
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
100 8
80
+ Y& + f
Control GR91669 O.l/.tM GR91669 0.3uM GR91669 l.O;M f
r
Control Echistatin 1OnM Echistatin 30nM Echistatin 1OOnM &&
b
80
Y/ Z$#&
‘Z
0
279
60
60
kh on 40 c 8 20
T
,
0 1
10
Fibrinogen
100 Concentration
1000
10000
@g/ml)
10
100
Fibrinogen
Concentration
1000
10000
@g/ml)
FIG. 5 The effects of GR91669 (lefthand panel) or echistatin (righthand panel) on fibrinogendependent aggregation of platelets prestimulated with ADP (10~M). Points are means and S.E.M of 4-7 experiments. Where not shown error bars are within the symbols.
Marmoset whole blood platelet aggregation. GR91669 and echistatin caused 90-100% inhibition of ADP (30pM)-induced platelet aggregation in marmoset whole blood, with IC50 (n=4) values of 0.5 (0.3-0.9)kM and 0.12 (0.05-0.27)~M respectively which are similar to those obtained in human whole blood (Table 1). GRGDS caused a maximum of 64% inhibition with an IC50 of 566 (3151015pM, n=3), which was approximately 4-fold greater than in human whole blood. -H- Control Y& 8 mins post dose + 15 mins post dose
1
I 10 ADP concentration
i
I 100 (PM)
+ U100 _+ 0
1
Control 60 mins post dose 120 mins post dose 240 mins post dose
10 ADP concentration
I 100 (PM)
FIG. 6
The effect of intravenous administration of lmgjkg (left-hand panel) or lOmg/kg (tight-hand panel) GR91669 on ADP-induced marmoset platelet aggregation ex vivo. Points are means f: s.e.m. of from 3-6 animals. Where not shown error bars are within the symbols.
280
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
Intravenous administration of lmg/kg or lOmg/kg GR91669 to conscious marmosets caused no untoward effects except for dispersed bruising at points of restraint in a single animal in each dose group. ADP (1-30uM)-induced platelet aggregation ex vivo was completely abolished in blood samples taken at 15 minutes (lmg/kg) and 1 hour (10mgkg) after dosing (Figure 6). Greater than 50% inhibition was maintained for 30 minutes and 2 hours respectively. ADP concentration-effect curves returned to within 2-4 -fold of control sensitivity at approximately 1 hour and 4 hours after lmg/kg and lOmg/kg GR91669. No significant inhibition of ADPinduced platelet aggregation ex vivo was observed after administration of vehicle.
DISCUSSION
We have previously shown that the potency of RGD-like molecules as inhibitors of platelet aggregation could be improved by introducing conformational restraints by cyclisation, as in GR83895 (5), to mimic the RGD-conformation in the a-chain of fibrinogen. In the search for more potent and specific fibrinogen receptor (GpIIb/IIIa) antagonists, a series of modified, unconstrained, linear peptides based upon RGDF were synthesised. Inhibitory activity was retained after the removal of the terminal amino group of arginine (compound I) suggesting little importance for it in binding to the fibrinogen receptor. However, the alkyl-guanidine side chain of arginine appears to be more important in receptor binding. Replacement of the guanidine with a simple amino group, lead to a slight loss in activity, but together with extension of the alkyl chain, produced a series of more readily prepared compounds. In contrast, replacement of guanidine by isothiourea caused 6-17 fold increases in potency. Furthermore, the distance between the N-terminal basic group and the carboxylic groups also affected activity. Previously, using nuclear magnetic resonance studies, it was shown that the distance between the arginine guanidine and aspartate carboxylate groups in cyclic peptide fibrinogen receptor antagonists was approximately 14A for optimal activity (4). Similarly, by varying the alkyl chain length in the current series of compounds it was found that a 6 carbon chain (compound V) resulted in optimal activity. This substitution also results in an approximate 14-16A distance between the primary amino group of arginine and carboxylic acid group of aspartate, as determined by molecular modelling of the compound’s linear extended conformation. This suggests that both these groups are involved in binding to one or more sites on GpIIb/IIIa and that the distance between them is critical. Finally, in GR91669, in which isothiourea replaced guanidine and both arginine and glycine were replaced by an alkyl chain, this 14-16A distance was maintained and produced the most potent compound of the series.
The findings described in this paper are consistent with GR91669 being a fibrinogen receptor antagonist. The compound inhibits platelet aggregation induced by all agonists studied, i.e. ADP, U46619, thrombin and collagen. Binding studies using 1251-fibrinogen demonstrated that GR91669 (1 MM) could totally prevent fibrinogen binding to ADP-stimulated platelets, with a potency similar to that which inhibited platelet aggregation. To rule out an inhibitory activity of GR91669 on stimulus-effect coupling in platelets, further studies were undertaken to assess its effects on platelet shape change and 14C-5HT secretion. GR91669 did not inhibit platelet shape change, whilst this could be completely abolished by prostacyclin. Similarly, GR91669 did not inhibit 14C-5HT secretion induced by U46619, collagen or thrombin. However, in PRP,
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
281
secretion induced by ADP was inhibited by GR91669, thus suggesting that this response may be secondary to fibrinogen binding to GpIIb/IIIa and subsequent aggregation. Whether this observation has any therapeutic implication is not clear, but any reduction in secretion from platelets during activation in vivo may be beneficial. Throughout the mechanistic studies, GR91669 demonstrated a profile similar to that of echistatin and GRGDS, providing further evidence that the compound acts as a fibrinogen receptor antagonist.
In competition studies, the inhibition of platelet aggregation caused by GR91669 or echistatin was reversed by increasing the concentration of fibrinogen. Therefore, perhaps not surprisingly, the fibrinogen receptor antagonists studied tended to be weaker in whole blood, where the concentration of fibrinogen is approximately 6-fold that in GFP (5OOpg/ml). In addition, the compounds tended to be weaker inhibitors of U46619- than ADP-induced platelet aggregation. Again, an increase in the fibrinogen concentration caused by the ability of U46619 to induce secretion of platelet fibrinogen may in part explain this difference, but in addition, expression of intracellular pools of GpIIb/IIIa would also contribute to a decrease in potency (9).
Since GpIIb/IIIa is a member of the integrin family, one important aspect of fibrinogen receptor antagonists as potential anti-thrombotic drugs is their specificity of action. The vitronectin receptor, avp3, bears a particularly close homology with GpIIb/IIIa, having the same P-subunit and sharing many common amino acid sequences within the a-subunits (10). There is evidence that the vitronectin receptor may contribute to the integrity of the adhesion of human endothelium to its underlying surface (11,12). Therefore, in vivo administration of a compound which failed to distinguish GpIIb/IIIa from the endothelial vitronectin receptor could lead to compromise of the integrity of the endothelial barrier and unwanted side-effects. We have previously shown (5), by the use of antibodies to the vitronectin receptor, that endothelial cells grown under the conditions described in this study adhere via a vitronectin receptor-dependent process. Echistatin and GRGDS both caused detachment of endothelial cells from culture wells, with potencies similar to those required to inhibit platelet aggregation. Detached endothelial cells remained viable, as shown by their ability to grow upon re-culture. In contrast, however, GR91669 had no effect on the attachment of endothelial cells, even at concentrations 500-fold greater than those required to inhibit platelet aggregation. In addition, GR91669 demonstrates a degree of specificity greater than previously shown for GR83895, which exhibited only a lOOfold difference in potency for fibrinogen over vitronectin receptors. Currently, it is not known which structural features of GR91669 has brought about such increased specificity. Studies were also undertaken on the adhesion of human granulocyte preparations to fibrinogen-coated plastic surfaces. Whilst there is little to suggest that RGD-containing peptides interfere with the immunological role of the complement binding integrin, Mac- 1, fibrinogen has been shown to interact with this molecule (13,14). Consequently, we have studied the adhesion of fMLPstimulated granulocytes to immobilised fibrinogen in a system, where antibodies to the Mac-l complex can completely inhibit the extent of deposition. In this in vitro system, none of the compounds tested had any inhibitory effect.
Previous reports have suggested that aggregation of platelets from several common laboratory animals is not sensitive to inhibition by fibrinogen receptor antagonists (6), and therefore these
282
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
species are unsuitable for in vivo evaluation of such compounds. However, we have demonstrated that GR91669 and echistatin inhibited marmoset platelet aggregation in whole blood in vitro, with potencies similar to those achieved in human whole blood. Marmosets were therefore chosen as a suitable species for ex vivo studies. Intravenous administration of GR91669 at lmg/kg or lOmg/kg produced a dose related effect. ADP-induced platelet aggregation ex vivo was abolished for 15 minutes and 60 minutes but these effects were reversible, with aggregation returning to control sensitivity over a period of 1 hour and 4 hours respectively. Compound V has been described previously, as SC-49992 (15), and reported to have a similarly short duration of action and short pharmacokinetic half life (approximately 10 min) in dogs. Likewise, Ro 43-5054 (16), another non-cyclic peptidic fibrinogen receptor antagonist, has also been reported to be short-acting after administration to dogs. These observations suggest that a long duration of action may not be easy to achieve with peptidic compounds. However, the short duration of action seen with this class of compound may be a useful profile in acute treatment of thrombotic disorders, as in the event of bleeding problems, termination of administration would rapidly restore platelet function. GR91669 would, however, be expected to be an effective antithrombotic in vivo, as both SC-49992 and Ro 43-5054, which have similar potency and specificity of action to GR91669, have been shown to prevent thrombus formation in canine models of coronary thrombosis (15,16).
In conclusion, GR91669 appears to be a specific inhibitor of the interaction of fibrinogen with platelet GpIIb/IIIa and inhibits aggregation induced by a range of agonists, including thrombin. Limited studies on the effects of intravenous administration of GR91669 to marmosets have shown it to have a short duration of action. This, and the lack of oral activity of other related peptidic fibrinogen receptor antagonists, highlights the need for non-peptidic compounds. Orally-active non-peptidic compounds, such as GR144053 (17,18,19), have now been reported. Acknowledgements We would like to thank Mr D Allen for his helpful discussions regarding the structure activity relationships, Mr A Crame and Mr A Pipe for the synthesis of the compounds and also Dr P Lumley for his help in preparing this manuscript. REFERENCES
1. SEISS, W. Molecular mechanism of platelet activation. Physiol Rev 69,58-178, 1989 2. OATES, J.A., FITZGERALD, G.A., BRANCH, R.A., JACKSON, E.K., KNAPP, H.R., ROBERTS, L.J. Clinical implication of prostaglandin and Thromboxane A formation. N. Engl. J. Med. 319,689-698,1988. 3. PYTELA, R., PIERSCHBACHER, M.D., GINSBERG, M.H., PLOW, E.F. and RUOSLAHTI, E. Platelet membrane glycoprotein IIb/IIIa: member of a family of Arg-Gly-Asp-specific receptors. Science. 231, 1559-1562, 1986.
adhesion
283
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75. No. 3
4. HANN, M.M., CARTER,
B., KITCHIN,
J., WARD, P., PIPE, A., BROOMHEAD,
J.,
HORNBY, E., FOSTER, M., and PERRY, C. An investigation of the bioactive conformation of Arg-Gly-Asp containing cyclic peptides and snake venom peptides which inhibit human platelet aggregation. Proceedings of the 2nd Molecular Recognition Conference. chemical and biochemical problems II ed. S.M.Roberts. pp 145160.
Molecular Recognition,
Royal Society Chemistry,
1992. 5. FOSTER, M.R., HORNBY, E.J., BROWN, S., KITCHIN, J., HANN, M., and WARD, P. Improved potency and specificity
of Arg-Gly-Asp
receptor blocking drugs. Thromb Res. 12,231-245, 6. HARFENIST,
E.J., PACKHAM,
(RGD) containing peptides as fibrinogen 1993.
M.A. and MUSTARD, J.F. Effects of the cell adhesion
peptide, Arg-Gly-Asp-Ser on responses of washed platelets from humans, rabbits and rats. Blood. 71, 132-136, 1988. 7. HORNBY, E.J., FOSTER, M.R., MCCABE, P.J., STRATTON, L.E. The inhibitory effects of GR32191, a thromboxane receptor blocking drug on human platelet aggregation, adhesion and secretion.
Thromb Haemostas 61,429-436,
8. SANTINI,
G.F., BATTISTIN,
1989.
S., DE PAOLI, P., VILLALTA,
D., BOREAN,
M. and
BASAGLIA, G. A method for the determination of the adherence of granulocytes to microtitre plates. J. Immunol. Methods 100, 153-156, 1987. 9. NIIYA, K., HODSON, E., BADER, R., BYERS-WARD, V., KOZIOL, J.A., PLOW, E.F. and RUGGERI, Z.M. Increased surface expression of the membrane glycoprotein IIb/IIIa complex induced by platelet activation. Blood. 70,475-483,
Relationship to binding of fibrinogen and platelet aggregation.
1987.
10. FITZGERALD,
L.A., PONCZ, M., STEINER,
B., RALE, S.C., BENNET,
J.S. and
PHILLIPS DR. Comparison of the cDNA derived sequences of the human vitronectin receptor alpha subunit and platelet glycoprotein IIb. Biochem J 26, 8158-8165, 1987. 11. KRAMER, R.H., CHENG. Y.F. and CLYMAN R. Human microvascular endothelial cells use beta- 1 and beta-3 integrin receptor complexes to attach to laminin. J Cell Biof. 111, 1233-1243, 1990. 12. ALBELDA,
S.M., DAISE, M., LEVINE,
E.M. and BUCK. C.A. Identification
and
characterisation of cell-substratum adhesion receptors on cultured human endothelial cells. J. Clin. Invest. 83, 1992-2002,
1992
13. ALTIERI, D,C., BADER, R., MANNUCI, P.M. and EDGINGTON, T.S. Oligospecificity
of the cellular receptor MAC-l encompasses an inducible recognition specificity for fibrinogen. J. Cell. Biol. 107, 1893-1900, 14. TREZZINI,
1988.
S., SCHEUPP, B., MALY, F.E. and JUNGI, T.W. Evidence that exposure to
fibrinogen or to antibodies directed against Mac- 1 (CD 11 b/CD1 8;CR3) modulates monocyte effector function. Br. J. Haematol. 77, 16-24, 1991.
human
284
FIBRINOGEN RECEPTOR ANTAGONISTS
15. FEIGEN, L.P., KING, L.W., TAITE, B.B., PANZER-KNODLE,
NICHOLSON,
N.S. Pharmacology
Vol. 75, No. 3
S.G., SALYERS, A.K.,
of SC-49992, a peptide mimetic that inhibits platelet Gp
IIb/IIIa. Thrombos Huemostas 65,8 12, Abstract 456, 199 1. 16. ROUX, S., STEINER, B., HADVARY, P. and WELLER, T. Selective blockade of the platelet receptor GpIIb-IIIa by a non-cyclic
peptidomimetic:
thrombus formation. Thrombos Huemostus 65,812,
Effects on in-vitro and in-vivo
Abstract 458,199l.
17. FOSTER, M.R., PIKE, N.B., HORNBY, E.J., PORTER, B., ELDRED, C.D., ROSS, B.C. and LUMLEY, P. GR144053, a potent and specific non-peptide fibrinogen receptor antagonist on human platelets in vitro. Thrombosis and Haemostas 69,559,
Abstract 64, 1993.
18. PIKE, N.B, FOSTER, M.R., HORNBY, E.J. and LUMLEY, P. Effect of the fibrinogen receptor antagonist GR144053 upon platelet aggregation ex vivo following intravenous and oral administration to the marmoset and cynomolgus monkey. Thrombosis and Haemostas 69, 1071, Abstract 1886, 1993. 19. WHITE, B.P., and LUMLEY, P. GR144053, a fibrinogen receptor antagonist, prevents coronary thrombosis in the anaesthetised dog. Thrombosis and Haemostas 69, 1071, Abstract 1884, 1993.
INHIBITION OF HUMAN PLATELET AGGREGATION BY GR91669, A PROTOTYPE FIBRINOGEN RECEPTOR ANTAGONIST Martyn R.Foster, Edward J.Hornby, Susan Brown, Michael Harm’, John Kitchinl, Nicholas Pike, and Peter Ward’. Departments of Cardiovascular and Respiratory Pharmacology and ‘Medicinal Chemistry,Glaxo Group Research Limited, Ware, Herts.
(Received 23 February 1994 by Editor M.E Scully; revised/accepted 5 May 1994)
Abstract
In order to produce more potent and specific fibrinogen receptor (GpIIb/IIIa) antagonists, the Arg-Gly of a chemical series based upon Arg-Gly-Asp was replaced by alkyl chains of varying lengths. The most potent in this series, GR91669, inhibited aggregation of human gel-filtered platelets (GFP) in vitro induced by ADP or the thromboxane A2 mimetic, U46619, with IC50 values of 200nM and 500nM respectively and was selected for further studies. Its inhibitory effects on GFP were reversed by addition of excess fibrinogen. The compound also inhibited ADP- or U466 19- induced latelet aggregation in human whole blood (IC50 values of 700nM in both cases). 12PI-Fibrinogen binding to ADP-stimulated platelets was inhibited by GR91669 with an IC50 ( 65nM) similar to that against platelet aggre ation. GR91669 (1mM) did not inhibit U46619-induced platelet shape change or 1PC-5HT secretion from platelets stimulated by collagen, U46619 or thrombin. Therefore GR91669 inhibits aggregation but has no significant effect on stimulus-response events, a profile consistent with fibrinogen receptor blockade. In addition, GR91669 (ImM), unlike echistatin or Gly-Arg-Gly-Asp-Ser, did not disrupt vitronectin receptor-dependent attachment of cultured HUVECS in vitro and similarly did not inhibit Mac-l dependent adhesion of human granulocytes. Thus, of the integrins tested, GR91669 appears to be specific for GpIIb/IIIa. Following intravenous administration to marmosets of 1 or lOmg/kg GR91669, ADP (lObM)-induced platelet aggregation ex vivo was abolished for 15 and 60 minutes respectively. Greater than 50% inhibition was maintained for 30 minutes and 2 hours respectively. GR91669, therefore appears to be a potent, specific fibrinogen receptor antagonist in vitro and which is also active in vivo.
Key words: GR91669, GR83895, echistatin, human platelets, Glycoprotein IIb/IIIa, Fibrinogen receptor antagonists, Integrins.
269
270
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
Platelets play a central role in thrombosis. After damage to the endothelial cell lining of blood vessels, proteins are exposed providing a surface on which platelets may adhere. Subsequently, adhered platelets become activated and through secretion of aggregatory agonists recruit further platelets to the injury site. During such activation, and as a result of exposure to aggregatory agonists, glycoprotein (Gp) IIb/IIIa, the fibrinogen receptor, changes its conformation and binds fibrinogen (1). Such binding results in the crosslinking of platelets by fibrinogen leading to aggregation and thrombus formation. This may result in obstructed blood flow with potentially fatal consequences.
Current anti-thrombotic therapies prevent the action of a single aggregatory agonist, but do not attenuate the effects of other agonists. For example, aspirin prevents the formation of thromboxane A2 but does not prevent aggregation induced by other agonists such as adenosine diphosphate (ADP) or thrombin (2). However, the binding of fibrinogen to GpIIb/IIIa represents a common event during platelet aggregation, regardless of the initiating stimulus. Thus, antagonism of platelet fibrinogen receptors inhibits platelet aggregation to a broad spectrum of aggregatory agonists, and represents a pharmacological approach more effective than current therapies to the potential treatment of thrombosis.
The amino acid sequence, Arg-Gly-Asp (RGD) is present in the Aa chain of fibrinogen and binds to GpIIb/IIIa (3). It is predicted that RGD exists in a p-loop conformation within fibrinogen (4). We have synthesised and previously reported results obtained with cyclicpeptides based upon RGD. GR83895, a synthetic cyclic peptide, causes inhibition of fibrinogen binding and platelet aggregation with improved potency compared with homologous linear peptides (5). GpIIb/IIIa is a member of the integrin super-family of adhesion receptors and the GpIIIa (p3) subunit is common to the fibrinogen, vitronectin and fibronectin receptors. Other fibrinogen receptor antagonists such as GRGDS and echistatin appear to have additional inhibitory effects on vitronectin receptors, whilst GR83895 is lOO-fold more specific for fibrinogen receptors (5). However, GR83895 lacks the potency of other fibrinogen receptor antagonists such as echistatin, and therefore we have synthesised a series of linear molecules in an attempt to obtain more potent fibrinogen receptor antagonists with similar specificity of action.
This study describes the in vitro structure activity relationships of a series of novel linear peptidic compounds based upon RGD, and the in vitro pharmacological profile and specificity of action of the most potent compound, GR91669. Previous reports have suggested that platelets from several commonly used laboratory species, including rat and rabbit, are not inhibited by fibrinogen receptor antagonists (6). After preliminary in vitro investigations, marmoset platelets were found to be similar to human platelets in sensitivity to inhibition by fibrinogen receptor antagonists, and therefore, we also describe the potency and duration of action of GR91669 after intravenous administration in this species.
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
271
MATERIALS AND METHODS
Materials
GR91669 (8-[(aminoiminomethyl)thio]-L-aspartyl-L-phenylalanine), GR83895 (5 ), compounds I-VIII (as trifluoroacetate salts) and prostacyclin were synthesised in the Chemistry Division, Glaxo Group Research. GR91669 was prepared as a 1mM stock solution in 5mM Hepes in 0.9% w/v sodium chloride solution (saline) (pH 8.4) and diluted with 5mM Hepes in saline (pH7.4) for use. Compounds I-VIII were dissolved in dimethylsulphoxide (DMSO) and diluted in 5mM Hepes buffered saline (pH 7.4) to give a stock concentration of 1mM in 10% v/v DMSO. GR83895, Gly-Arg-Gly-Asp-Ser (GRGDS; Bachem AG, Bubendorf, West Germany) and echistatin (Bachem (UK) Ltd) were prepared as 1mM stock solutions in 5mM Hepes buffered saline (pH 7.4). Prostacyclin was prepared as a 1mM stock solution in 50mM Tris buffer (pH 9.0), stored at -20°C and was diluted to working concentrations in 50mM Tris buffer (4’ C, pH 8.0) on the day of use. Aspirin (Sigma Chemical Co.) was prepared as a lOmg/ml solution in 1OOmM Tris buffer (pH 8.5). Adenosine diphosphate (ADP; Sigma Chemical Co.) was dissolved in 50mM Tris buffer (pH 6.5) as a 1mM stock solution. U46619, a chemically stable TxA2 mimetic (Cayman Chemicals) was prepared as a 1mM stock solution in 1% w/v NaHC03 in saline. Equine collagen suspension was from Hormon Chemie, Munich, West Germany. Human cr-thrombin (Sigma Chemical Co.) was prepared as a 10 IU/ml solution in saline. Ethylene glycolbisaminoethylether tetra-acetic acid (EGTA) was prepared as a 400mM stock solution in 20mM Hepes buffered saline (pH7.4) and adjusted to pH7.4 by addition of 5N NaOH. yF5rnan fibrinogen was from Kabt.’ 14C-5-hydroxytryptamine (14C-5HT; SOmCi/mmol) and Iodine were from Amersham International Plc. Penicillin, gentamycin, streptomycin, Hank’s medium, medium 199 (M199), foetal calf serum and glutamine were from ICN-Flow, High Wycombe, UK. Collagenase was from Boehringer Mannheim, Lewes, UK. Rose Bengal stain was from Koch-Light Ltd, Haverhill, UK. Monoclonal antibodies to CD1 1 (anti-Macl) and GpIIIa were from Dakopatts, Denmark. Hank’s buffered salt solutions (HBSS) and sterile Hepes, bovine serum albumin and sodium bicarbonate solutions were from Gibco. All other reagents used during the preparation of granulocytes were from Sigma Chemical Company. All other reagents were from BDH Ltd., Poole and were of analar grade.
Platelet preparation
Blood from human volunteers who had taken no medication for 10 days was collected by venepuncture into anticoagulant (9 vol. blood: 1 vol. 0.13M trisodium citrate solution). Plateletrich plasma (PRP) and gel-filtered platelets (GFP) were prepared as described previously (7). Fibrinogen was added to GFP to a final concentration of 5OO&ml (5) except in experiments where competition studies were performed.
Platelet aggregation
Aggregation of GFP was induced by addition of ADP (10pM) and studied using a 6 channel, optical aggregometer (Glaxo Bioengineering Unit) as described previously (5). Increasing concentrations of potential inhibitors of aggregation were incubated with GFP for 5 minutes at 37 OC, prior to addition of ADP. Aggregation was expressed as the percentage change in light
272
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
transmittance compared with a vehicle-treated, aggregated sample. IC50 values were calculated as the concentration of compound which caused 50% inhibition of aggregation.
Platelet aggregation in human titrated, whole blood was quantified by counting the fall in single platelet count after addition of U46619 (3pM) or ADP (301rM), according to a method described previously (5). IC50 values were calculated for each of the compounds tested, as the concentration of inhibitor which caused a 50% inhibition of the fall in platelet count measured.
12jI-Fibninogen binding studies. GRGDS, GR91669 and echistatin were tested as inhibitors of 1251-fibrinogen (2pCi, 5OO&ml final cont.) binding to ADP (20pM -stimulated GFP, according to the method described previously (5). Bound and unbound 1251-fibrinogen were separated by centrifugation through dioctyl phthalate / dibutyl phthalate (6:4 v/v) mixtures and the radioactivity in the platelet pellets estimated by scintillation counting.
Plateletshape change Platelet shape change was measured optically in a 6-channel aggregometer as the decrease in light transmittance through a sample of aspirin-treated PRP. By amplifying the output from the aggregometer it was possible to record shape change. Platelet aggregation was prevented by addition of EGTA (4mM) lmin prior to the induction of platelet shape change with U46619 (0.3~M). Compounds were incubated with aliquots of stirred (90Orpm) PRP (37’C) for 5 mins before agonist addition. The peak decrease in light transmittance was measured and expressed as a percentage of that obtained in a vehicle-treated sample.
14C-5HT secretion 14C-5HT secretion was measured in PRP or GFP prepared from blood, incubated (37°C 30 min) with 14C-5HT (lccCi/20ml), according to the methods described previously (7). GR91669 or prostacyclin concentration-inhibition curves were constructed in PRP against platelet aggregation and concurrent l4 C-5HT secretion induced by U46619 (lOpM), collagen (2rrglml) or ADP (10pM). Similarly, concentration-inhibition curves for GR91669, prostacyclin or echistatin were constructed in GFP against thrombin (O.lIU/ml)-induced platelet aggregation and 14C-5HT secretion. In order to avoid thrombin-induced clotting, fibrinogen was omitted from the suspension medium of GFP. 14C-5HT secretion was expressed as a percentage of the total 14C5HT present in an equal volume of PRP.
Competitionstudies Six cuvettes containing GFP, prepared without addition of fibrinogen, were incubated without stirring for 5 min at 37’ C, with a single concentration of inhibitor or vehicle. A different concentration of fibrinogen over the range l-2000~g/ml was then added to each cuvette and stirring (900 rpm) initiated. After 1 min, ADP (lOtiM) was added and the maximum aggregation occurring within 10 min was measured. A series of such aggregation curves was generated in a
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
randomised order, with up to three different concentrations way in each experiment.
273
of inhibitor examined in a similar
Detachment of human umbilical vein endothelial cells (HUVECs) Inhibitors were tested for their ability to detach HUVECs according to the method described previously (5). Compounds, prepared in culture medium, were incubated (2 hours, 37°C) with HUVEC cultures grown to confluence in 24 well plates. After washing the wells with fresh culture medium, the density of the remaining cells was estimated using a staining technique (5) and the results expressed as percentage detachment relative to wells containing HUVECs, which had not been exposed to compounds. In separate experiments, cultured HUVECs grown to confluence, were treated with GRGDS (1mM) or echistatin (1 BM) for 2 hours (37’C). Detached cells were harvested and washed in culture medium, before plating onto fresh culture plates and incubated (37°C under 5% v/v C02) until confluent.
Granulocyte adhesion Human whole blood was collected into anticoagulant (130mM trisodium citrate, 1mM sodium dihydrogen orthophosphate, 1lmM glucose; 6.7 vol. blood : 1 vol. anticoagulant). Red blood cells were removed by sedimentation through methyl cellulose (40m1, 13% w/v) containing sodium diatrizoate (0.6% w/v). The plasma fraction containing the granulocytes was retained and subsequently the granulocytes purified by centrifugation (4OOxg, 20 mins) on Ficoll-Paque. Contaminating red cells were removed by lysing with 0.2% w/v sodium chloride. After addition of 1.6% w/v sodium chloride to restore isotonic balance, the granulocytes were washed by centrifugation and resuspension in calcium free-Hank’s buffered salts solution. The granulocyte concentration was determined using a Coulter JT haematology analyser and adjusted to give a final concentration of 1~10~ cells /ml.
Purified human fibrinogen (300&ml) was dispensed into 24 well plates and incubated (17-20 hrs, 4°C) prior to use. The fibrinogen solution was replaced with HBSS containing 0.1% bovine serum albumin and incubated for lhr at 20°C. Each well was aspirated, test compound or vehicle (100~1) added and equilibrated (37°C) in a Dynatech shaker-incubator prior to the addition of granulocyte suspension (lml). After mixing for 10 mins, formyl-met-leu-phe (fMLP) (5nM) was added and incubated for 30mins (37°C) without shaking. Plates were agitated for lmin, each well aspirated, and lml of HBSS added and the procedure repeated. Adhered granulocytes were lysed by addition of cetyltrimethylammonium bromide (CTAB 0.3% w/v, 250~1; 30mins) and the lysate assayed for alkaline phosphatase, according to the method of Santini et al (8). Alkaline phosphatase values were expressed as a percentage of that obtained from a lml sample of cell suspension.
Marmoset whole bloodplatelet aggregation. Blood was collected by venepuncture from the femoral artery of conscious, restrained marmosets Callithrix jacchus (300-500gm) of either sex and anticoagulated as for human blood. Aliquots (150~1) of blood were stirred in an aggmgometer (lOOrpm, 37°C) and a sample (10~1) taken from each for control platelet counting using an Ultra-F10 100 platelet counter (Becton Dickinson).
274
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
After 5 minutes incubation (37’C) in the presence of GR91669 (O.l-lOtiM), echistatin (0.03IrrM), GRGDS (lo-1OOOpM) or vehicle, aggregation was induced by addition of a justmaximally effective concentration of ADP (30pM), and a further sample taken for platelet counting 1 minute after agonist addition. Prior to platelet counting, the (10~1) blood samples were added to 2ml of formal-saline (0.2% w/v), containing ethylene-diamine-tetraacetic acid (EDTA; 1mM). The percentage fall in platelet number was calculated relative to the control platelet count prior to addition of agonist. Inhibition of aggregation was expressed by comparing the percentage fall in platelet number, with that obtained in a maximally aggregated (30pM ADP) sample.
EX vivo whole blood platelet aggregation was performed in a similar manner. GR91669 (1 or lOmg/kg) or vehicle was administered to marmosets (300-5OOg) of either sex by intravenous injection, via a hind leg vein. Blood (0.6ml) was collected from a femoral vein, before, and at intervals up to 4 hours after administration of compound. Aliquots of blood (150~1) were dispensed into four cuvettes and ADP (l-30pM) concentration-effect curves of aggregation constructed as described above. RESULTS
Structure-activitystudies. We have previously shown that the potency of RGD-like molecules as inhibitors of platelet aggregation could be improved by cyclisation (5). In the search for more potent and specific fibrinogen receptor antagonists a series of linear peptides based upon RGDF were synthesised. When tested in GFP or whole blood, all the peptides caused complete inhibition of platelet aggregation induced by ADP (10pM) or U46619 (3pM). GRGDS was a relatively weak inhibitor, particularly when measurements were made in whole blood (Table 1). Removal of the N-terminal glycine and substitution of serine with phenylalanine gave RGDF which was 3-fold more potent as an inhibitor of ADP-induced platelet aggregation in GFP (IC50 8pM; Table 1). In an attempt to achieve further increases in potency, modifications were made to the linear RGDF sequence. The assumption that the interaction of the arginine with the fibrinogen receptor was through its alkyl-guanidine side chain was supported by the observation that activity was retained in compound I (ICso 5.1 bM), in which the N-terminal amino group of arginine was removed, but alkyl-guanidine retained (Figure 1). Further modification of compound I showed that the guanidine could be replaced by a simple primary amino group (compound II; IC50 14gM), with only a 2-3-fold loss in potency. In this series of more readily prepared compounds, we investigated the effect of different alkyl chain lengths (3-7 carbon atoms), thereby varying the distance between the terminal amino and carboxylic acid groups. As can be seen from Figure I (compounds II-VI), a chain length of 6 carbons produced optimal activity (compound V, IC50 1.8pM). Having optimised the alkyl chain length, the primary amino group in compound V was replaced with isothiourea, a group which more closely resembles a guanidine function, in order to regain the activity lost in compound II. However, to maintain the optimal length of compound V it was necessary to shorten the alkyl chain from 6 to 4 carbons, thus compensating for the additional length of isothiourea. These changes resulted in a 40-fold increase in potency (compound VII, IC50 0.3pM) over that of compound I. Having succesfully replaced the arginine of RGDF and enhanced potency we turned our attention to the possibility of replacing glycine.
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
275
Glycine often acts only as a peptide spacer and therefore we recognised that again it might be replaced by a simple non-peptidic alkyl chain spacer. Replacement of the glycine peptidic bond of compound I by an alkyl chain gave compound VIII, which had a 4-fold increase in potency (IC50 1.2~M). Finally, the guanidine group of compound VIII was again replaced by an isothiourea to give GR91669 (IC50 0.2~M), the most potent compound of this series. In comparison with the highly potent snake venom peptide, echistatin, GR91669 was 2-4 times less potent in GFP and 3-9 times less potent in whole blood (Table 1). In whole blood, the rank order of potency for the compounds tested was maintained, but with a tendency for the weaker compounds to decrease slightly in potency against a given aggregatory agonist, relative to GFP suspensions (Table 1).
X-(CH&CO-Gly-Asp-Phe Compound
X
IC50 (ctM)vs ADP
I
Guanidine
5.1
II
14.0
III
45.0
IV
64.0
V
1.8
VI
3.0
VII
0.3
Isothiourea
Y-(CH$,CO-Asp-Phe Compound
Y
n
IQ0 (PM)
VIII
Guanidine
7
1.2
GR91669
Isothiourea
7
0.2
FIG. 1 Structures of compounds I-VIII and GR91669 related to inhibitory potency.
Fibrinogen binding studies. Each of the inhibitors tested prevented the binding of fibrinogen to ADP (lObM)-stimulated platelets (Figure 2). Echistatin and GR91669 had IC50 values of 0.05~M and 0.06bM
276
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
respectively. GRGDS was significantly less active with an IC50 of approximately 2OrrM, although complete inhibition of binding could be achieved at concentrations of 30Or1M. 100 g
go-
2 ho_ ._ A7 : 408
20o.OOl .Ol .l 1 10 Concentration (FM)
100
1000
FIG. 2 The effect of GR91669 (O,O), GRGDS (QH) or echistatin (A,A) on the ADP(lOrM)induced binding of 1251-fibrinogento human platelets (solid symbols) and aggregation (open symbols) in GPP. Points are mean and S.E.M. of at least 3 independent observations. Where not shown error bars arc within the symbols.
TABLE I The effect of a range of compounds on human platelet aggregation induced in GPP or blood after addition of ADP or U46619. Compound
GFP IC50 (PM) ADP(lOuM)
U46619(3pM)
ADP(30pM)
U46619(3fiM)
(1:;6) NT
132 (79-222) NT
372 (206-687) NT
GRGDS RGDF
v VII VIII
(&I) 8
NT
(O.lY4.0, 0.3 (0.1-1.6) 1.3 (1.0-1.7)
NT
(O.Z.3) 0.05 (0.04-0.06)
(O.Z7) 0.3 (0.06- 1.6)
NT
GR91669 Echistatin
Whole blood IC50 (PM)
3.1 (1.4-7.1) (OE.3) 5.9 (3.5-9.9) 0.7 (0.6-1.0) 0.08 (0.05-O. 12)
(l&3) (3.7Y7.8) 12.3 (6.7-22.9) (o.:~:.o) 0.2 (0.2-0.3)
Values are Geometric means and 95% confidence limits on at least 3 separate determinations. NT-not tested.
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
(b)
277
(4
FIG. 3
Typical aggregometer traces illustrating (a) shape change induced by U46619, (b) effects of GR91669 (1mM) and (c) prostacyclin (0.1~M) on U46619 (0.3rcM)-induced shape change in the presence of EGTA (4mM). Shape change studies
None of the inhibitors of platelet aggregation caused shape change on addition to PRP. Additionally, shape change induced by U46619 (0.3ctM) was not inhibited by any of the peptide analogues even at concentrations up to 1mM. In contrast, prostacyclin was potent in inhibiting shape change, with an IC50 value of 0.04pM (Figure 3). 14C-SHT secretion
U46619 (lOlM), collagen (2&ml) or ADP (10uM) induced platelet aggregation and concurrent secretion of 14C-5HT from stirred PRP (Figure 4). U46619- and collagen-induced secretion was not inhibited by GR91669 (O.l-3.0pM; Figure 4b) whilst, in contrast, ADP-induced secretion was totally inhibited. In the same PRP samples, GR91669 (0.3-3.0 PM) caused 90-100% inhibition of the concommitantly measured platelet aggregation induced by each of the agonists (Figure 4a). Both GR91669 (0.01-I PM) and echistatin inhibited thrombin-induced aggregation, but not 14C-5HT secretion, in GFP, whereas prostacyclin (lOnM-1OOnM) inhibited aggregation and secretion to all agonists in parallel, as illustrated against thrombin (Figure 4c & 4d).
Competition studies.
When added to ADP (lOpM)-stimulated platelets, fibrinogen caused a concentration-related increase in platelet aggregation as measured by optical aggregometery. GR91669 and GR83895 caused concentration-related, parallel, rightward shifts of fibrinogen concentration-effect curves (Figure 5). Echistatin at 0.1 PM caused an apparently unsurmountable inhibition of aggregation. It is not known whether higher concentrarions of fibrinogen might have restored the aggregation response.
Specificity studies
GR91669 did not cause detachment of HUVECs at concentrations up to 1mM (data not shown). Echistatin and GRGDS were both able to cause complete detachment of the cells at
278
(a)
(b) 100
+ + *
U46619 (10pM) ADP (10~M) Collagen (2pg/ml)
1
.l
r
Concentration GR91669 (FM)
(4
w
-
-& +
20 0
Prostacyclin CR91669
.g H $
80
g 3 ::
40
60
20 I
J
t
0 1
.OOl
.Ol
10
1
.l
10
Concentration GR91669 (FM)
e
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
I
I
.l
1
L I
.OOl
.Ol
I
I
.l
1
Inhibitor Concentration (FM)
Inhibitor Concentration (PM)
FIG. 4 The effect of GR91669 on aggregation and concurrently measured 14C-5HT secretion induced by U46619 (lObM), ADP(lOkM), or collagen (2&ml) in PRP (a Jr b) and induced by human thrombin (0.1 W/ml) in GFP (c & d). The results are the mean and S.E.M of 4 experiments. Error bars omitted in figure 4c for clarity. concentrations of 0.1 crM and 100~M respectively, with IQ values of 0.01 BM and 48pM (data not shown). Approximately 40% of cells remained attached in the presence of GR83895 (100~M). Cells which had been displaced by echistatin or GRGDS were harvested, washed and grew to confluence within a normal time scale, suggesting that detachment was not due to intrinsic cytotoxicity in either peptide. Granulocyte adhesion Of the granulocytes added to each well, approximately 20-40% became adherent after 30 min of incubation. None of the inhibitors tested caused any reduction in fMLP-stimulated granulocyte adhesion (data not shown). A monoclonal antibody to Mac-l (l/1000-l/10 v/v) caused concentration related reductions in granulocyte adhesion, with complete abolition at titres of l/10 v/v (data not shown). A monoclonal antibody of the same immunoglobulin class as that to Mac1 had no inhibitory effect.
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
100 8
80
+ Y& + f
Control GR91669 O.l/.tM GR91669 0.3uM GR91669 l.O;M f
r
Control Echistatin 1OnM Echistatin 30nM Echistatin 1OOnM &&
b
80
Y/ Z$#&
‘Z
0
279
60
60
kh on 40 c 8 20
T
,
0 1
10
Fibrinogen
100 Concentration
1000
10000
@g/ml)
10
100
Fibrinogen
Concentration
1000
10000
@g/ml)
FIG. 5 The effects of GR91669 (lefthand panel) or echistatin (righthand panel) on fibrinogendependent aggregation of platelets prestimulated with ADP (10~M). Points are means and S.E.M of 4-7 experiments. Where not shown error bars are within the symbols.
Marmoset whole blood platelet aggregation. GR91669 and echistatin caused 90-100% inhibition of ADP (30pM)-induced platelet aggregation in marmoset whole blood, with IC50 (n=4) values of 0.5 (0.3-0.9)kM and 0.12 (0.05-0.27)~M respectively which are similar to those obtained in human whole blood (Table 1). GRGDS caused a maximum of 64% inhibition with an IC50 of 566 (3151015pM, n=3), which was approximately 4-fold greater than in human whole blood. -H- Control Y& 8 mins post dose + 15 mins post dose
1
I 10 ADP concentration
i
I 100 (PM)
+ U100 _+ 0
1
Control 60 mins post dose 120 mins post dose 240 mins post dose
10 ADP concentration
I 100 (PM)
FIG. 6
The effect of intravenous administration of lmgjkg (left-hand panel) or lOmg/kg (tight-hand panel) GR91669 on ADP-induced marmoset platelet aggregation ex vivo. Points are means f: s.e.m. of from 3-6 animals. Where not shown error bars are within the symbols.
280
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
Intravenous administration of lmg/kg or lOmg/kg GR91669 to conscious marmosets caused no untoward effects except for dispersed bruising at points of restraint in a single animal in each dose group. ADP (1-30uM)-induced platelet aggregation ex vivo was completely abolished in blood samples taken at 15 minutes (lmg/kg) and 1 hour (10mgkg) after dosing (Figure 6). Greater than 50% inhibition was maintained for 30 minutes and 2 hours respectively. ADP concentration-effect curves returned to within 2-4 -fold of control sensitivity at approximately 1 hour and 4 hours after lmg/kg and lOmg/kg GR91669. No significant inhibition of ADPinduced platelet aggregation ex vivo was observed after administration of vehicle.
DISCUSSION
We have previously shown that the potency of RGD-like molecules as inhibitors of platelet aggregation could be improved by introducing conformational restraints by cyclisation, as in GR83895 (5), to mimic the RGD-conformation in the a-chain of fibrinogen. In the search for more potent and specific fibrinogen receptor (GpIIb/IIIa) antagonists, a series of modified, unconstrained, linear peptides based upon RGDF were synthesised. Inhibitory activity was retained after the removal of the terminal amino group of arginine (compound I) suggesting little importance for it in binding to the fibrinogen receptor. However, the alkyl-guanidine side chain of arginine appears to be more important in receptor binding. Replacement of the guanidine with a simple amino group, lead to a slight loss in activity, but together with extension of the alkyl chain, produced a series of more readily prepared compounds. In contrast, replacement of guanidine by isothiourea caused 6-17 fold increases in potency. Furthermore, the distance between the N-terminal basic group and the carboxylic groups also affected activity. Previously, using nuclear magnetic resonance studies, it was shown that the distance between the arginine guanidine and aspartate carboxylate groups in cyclic peptide fibrinogen receptor antagonists was approximately 14A for optimal activity (4). Similarly, by varying the alkyl chain length in the current series of compounds it was found that a 6 carbon chain (compound V) resulted in optimal activity. This substitution also results in an approximate 14-16A distance between the primary amino group of arginine and carboxylic acid group of aspartate, as determined by molecular modelling of the compound’s linear extended conformation. This suggests that both these groups are involved in binding to one or more sites on GpIIb/IIIa and that the distance between them is critical. Finally, in GR91669, in which isothiourea replaced guanidine and both arginine and glycine were replaced by an alkyl chain, this 14-16A distance was maintained and produced the most potent compound of the series.
The findings described in this paper are consistent with GR91669 being a fibrinogen receptor antagonist. The compound inhibits platelet aggregation induced by all agonists studied, i.e. ADP, U46619, thrombin and collagen. Binding studies using 1251-fibrinogen demonstrated that GR91669 (1 MM) could totally prevent fibrinogen binding to ADP-stimulated platelets, with a potency similar to that which inhibited platelet aggregation. To rule out an inhibitory activity of GR91669 on stimulus-effect coupling in platelets, further studies were undertaken to assess its effects on platelet shape change and 14C-5HT secretion. GR91669 did not inhibit platelet shape change, whilst this could be completely abolished by prostacyclin. Similarly, GR91669 did not inhibit 14C-5HT secretion induced by U46619, collagen or thrombin. However, in PRP,
Vol. 75, No. 3
FIBRINOGEN RECEPTOR ANTAGONISTS
281
secretion induced by ADP was inhibited by GR91669, thus suggesting that this response may be secondary to fibrinogen binding to GpIIb/IIIa and subsequent aggregation. Whether this observation has any therapeutic implication is not clear, but any reduction in secretion from platelets during activation in vivo may be beneficial. Throughout the mechanistic studies, GR91669 demonstrated a profile similar to that of echistatin and GRGDS, providing further evidence that the compound acts as a fibrinogen receptor antagonist.
In competition studies, the inhibition of platelet aggregation caused by GR91669 or echistatin was reversed by increasing the concentration of fibrinogen. Therefore, perhaps not surprisingly, the fibrinogen receptor antagonists studied tended to be weaker in whole blood, where the concentration of fibrinogen is approximately 6-fold that in GFP (5OOpg/ml). In addition, the compounds tended to be weaker inhibitors of U46619- than ADP-induced platelet aggregation. Again, an increase in the fibrinogen concentration caused by the ability of U46619 to induce secretion of platelet fibrinogen may in part explain this difference, but in addition, expression of intracellular pools of GpIIb/IIIa would also contribute to a decrease in potency (9).
Since GpIIb/IIIa is a member of the integrin family, one important aspect of fibrinogen receptor antagonists as potential anti-thrombotic drugs is their specificity of action. The vitronectin receptor, avp3, bears a particularly close homology with GpIIb/IIIa, having the same P-subunit and sharing many common amino acid sequences within the a-subunits (10). There is evidence that the vitronectin receptor may contribute to the integrity of the adhesion of human endothelium to its underlying surface (11,12). Therefore, in vivo administration of a compound which failed to distinguish GpIIb/IIIa from the endothelial vitronectin receptor could lead to compromise of the integrity of the endothelial barrier and unwanted side-effects. We have previously shown (5), by the use of antibodies to the vitronectin receptor, that endothelial cells grown under the conditions described in this study adhere via a vitronectin receptor-dependent process. Echistatin and GRGDS both caused detachment of endothelial cells from culture wells, with potencies similar to those required to inhibit platelet aggregation. Detached endothelial cells remained viable, as shown by their ability to grow upon re-culture. In contrast, however, GR91669 had no effect on the attachment of endothelial cells, even at concentrations 500-fold greater than those required to inhibit platelet aggregation. In addition, GR91669 demonstrates a degree of specificity greater than previously shown for GR83895, which exhibited only a lOOfold difference in potency for fibrinogen over vitronectin receptors. Currently, it is not known which structural features of GR91669 has brought about such increased specificity. Studies were also undertaken on the adhesion of human granulocyte preparations to fibrinogen-coated plastic surfaces. Whilst there is little to suggest that RGD-containing peptides interfere with the immunological role of the complement binding integrin, Mac- 1, fibrinogen has been shown to interact with this molecule (13,14). Consequently, we have studied the adhesion of fMLPstimulated granulocytes to immobilised fibrinogen in a system, where antibodies to the Mac-l complex can completely inhibit the extent of deposition. In this in vitro system, none of the compounds tested had any inhibitory effect.
Previous reports have suggested that aggregation of platelets from several common laboratory animals is not sensitive to inhibition by fibrinogen receptor antagonists (6), and therefore these
282
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75, No. 3
species are unsuitable for in vivo evaluation of such compounds. However, we have demonstrated that GR91669 and echistatin inhibited marmoset platelet aggregation in whole blood in vitro, with potencies similar to those achieved in human whole blood. Marmosets were therefore chosen as a suitable species for ex vivo studies. Intravenous administration of GR91669 at lmg/kg or lOmg/kg produced a dose related effect. ADP-induced platelet aggregation ex vivo was abolished for 15 minutes and 60 minutes but these effects were reversible, with aggregation returning to control sensitivity over a period of 1 hour and 4 hours respectively. Compound V has been described previously, as SC-49992 (15), and reported to have a similarly short duration of action and short pharmacokinetic half life (approximately 10 min) in dogs. Likewise, Ro 43-5054 (16), another non-cyclic peptidic fibrinogen receptor antagonist, has also been reported to be short-acting after administration to dogs. These observations suggest that a long duration of action may not be easy to achieve with peptidic compounds. However, the short duration of action seen with this class of compound may be a useful profile in acute treatment of thrombotic disorders, as in the event of bleeding problems, termination of administration would rapidly restore platelet function. GR91669 would, however, be expected to be an effective antithrombotic in vivo, as both SC-49992 and Ro 43-5054, which have similar potency and specificity of action to GR91669, have been shown to prevent thrombus formation in canine models of coronary thrombosis (15,16).
In conclusion, GR91669 appears to be a specific inhibitor of the interaction of fibrinogen with platelet GpIIb/IIIa and inhibits aggregation induced by a range of agonists, including thrombin. Limited studies on the effects of intravenous administration of GR91669 to marmosets have shown it to have a short duration of action. This, and the lack of oral activity of other related peptidic fibrinogen receptor antagonists, highlights the need for non-peptidic compounds. Orally-active non-peptidic compounds, such as GR144053 (17,18,19), have now been reported. Acknowledgements We would like to thank Mr D Allen for his helpful discussions regarding the structure activity relationships, Mr A Crame and Mr A Pipe for the synthesis of the compounds and also Dr P Lumley for his help in preparing this manuscript. REFERENCES
1. SEISS, W. Molecular mechanism of platelet activation. Physiol Rev 69,58-178, 1989 2. OATES, J.A., FITZGERALD, G.A., BRANCH, R.A., JACKSON, E.K., KNAPP, H.R., ROBERTS, L.J. Clinical implication of prostaglandin and Thromboxane A formation. N. Engl. J. Med. 319,689-698,1988. 3. PYTELA, R., PIERSCHBACHER, M.D., GINSBERG, M.H., PLOW, E.F. and RUOSLAHTI, E. Platelet membrane glycoprotein IIb/IIIa: member of a family of Arg-Gly-Asp-specific receptors. Science. 231, 1559-1562, 1986.
adhesion
283
FIBRINOGEN RECEPTOR ANTAGONISTS
Vol. 75. No. 3
4. HANN, M.M., CARTER,
B., KITCHIN,
J., WARD, P., PIPE, A., BROOMHEAD,
J.,
HORNBY, E., FOSTER, M., and PERRY, C. An investigation of the bioactive conformation of Arg-Gly-Asp containing cyclic peptides and snake venom peptides which inhibit human platelet aggregation. Proceedings of the 2nd Molecular Recognition Conference. chemical and biochemical problems II ed. S.M.Roberts. pp 145160.
Molecular Recognition,
Royal Society Chemistry,
1992. 5. FOSTER, M.R., HORNBY, E.J., BROWN, S., KITCHIN, J., HANN, M., and WARD, P. Improved potency and specificity
of Arg-Gly-Asp
receptor blocking drugs. Thromb Res. 12,231-245, 6. HARFENIST,
E.J., PACKHAM,
(RGD) containing peptides as fibrinogen 1993.
M.A. and MUSTARD, J.F. Effects of the cell adhesion
peptide, Arg-Gly-Asp-Ser on responses of washed platelets from humans, rabbits and rats. Blood. 71, 132-136, 1988. 7. HORNBY, E.J., FOSTER, M.R., MCCABE, P.J., STRATTON, L.E. The inhibitory effects of GR32191, a thromboxane receptor blocking drug on human platelet aggregation, adhesion and secretion.
Thromb Haemostas 61,429-436,
8. SANTINI,
G.F., BATTISTIN,
1989.
S., DE PAOLI, P., VILLALTA,
D., BOREAN,
M. and
BASAGLIA, G. A method for the determination of the adherence of granulocytes to microtitre plates. J. Immunol. Methods 100, 153-156, 1987. 9. NIIYA, K., HODSON, E., BADER, R., BYERS-WARD, V., KOZIOL, J.A., PLOW, E.F. and RUGGERI, Z.M. Increased surface expression of the membrane glycoprotein IIb/IIIa complex induced by platelet activation. Blood. 70,475-483,
Relationship to binding of fibrinogen and platelet aggregation.
1987.
10. FITZGERALD,
L.A., PONCZ, M., STEINER,
B., RALE, S.C., BENNET,
J.S. and
PHILLIPS DR. Comparison of the cDNA derived sequences of the human vitronectin receptor alpha subunit and platelet glycoprotein IIb. Biochem J 26, 8158-8165, 1987. 11. KRAMER, R.H., CHENG. Y.F. and CLYMAN R. Human microvascular endothelial cells use beta- 1 and beta-3 integrin receptor complexes to attach to laminin. J Cell Biof. 111, 1233-1243, 1990. 12. ALBELDA,
S.M., DAISE, M., LEVINE,
E.M. and BUCK. C.A. Identification
and
characterisation of cell-substratum adhesion receptors on cultured human endothelial cells. J. Clin. Invest. 83, 1992-2002,
1992
13. ALTIERI, D,C., BADER, R., MANNUCI, P.M. and EDGINGTON, T.S. Oligospecificity
of the cellular receptor MAC-l encompasses an inducible recognition specificity for fibrinogen. J. Cell. Biol. 107, 1893-1900, 14. TREZZINI,
1988.
S., SCHEUPP, B., MALY, F.E. and JUNGI, T.W. Evidence that exposure to
fibrinogen or to antibodies directed against Mac- 1 (CD 11 b/CD1 8;CR3) modulates monocyte effector function. Br. J. Haematol. 77, 16-24, 1991.
human
284
FIBRINOGEN RECEPTOR ANTAGONISTS
15. FEIGEN, L.P., KING, L.W., TAITE, B.B., PANZER-KNODLE,
NICHOLSON,
N.S. Pharmacology
Vol. 75, No. 3
S.G., SALYERS, A.K.,
of SC-49992, a peptide mimetic that inhibits platelet Gp
IIb/IIIa. Thrombos Huemostas 65,8 12, Abstract 456, 199 1. 16. ROUX, S., STEINER, B., HADVARY, P. and WELLER, T. Selective blockade of the platelet receptor GpIIb-IIIa by a non-cyclic
peptidomimetic:
thrombus formation. Thrombos Huemostus 65,812,
Effects on in-vitro and in-vivo
Abstract 458,199l.
17. FOSTER, M.R., PIKE, N.B., HORNBY, E.J., PORTER, B., ELDRED, C.D., ROSS, B.C. and LUMLEY, P. GR144053, a potent and specific non-peptide fibrinogen receptor antagonist on human platelets in vitro. Thrombosis and Haemostas 69,559,
Abstract 64, 1993.
18. PIKE, N.B, FOSTER, M.R., HORNBY, E.J. and LUMLEY, P. Effect of the fibrinogen receptor antagonist GR144053 upon platelet aggregation ex vivo following intravenous and oral administration to the marmoset and cynomolgus monkey. Thrombosis and Haemostas 69, 1071, Abstract 1886, 1993. 19. WHITE, B.P., and LUMLEY, P. GR144053, a fibrinogen receptor antagonist, prevents coronary thrombosis in the anaesthetised dog. Thrombosis and Haemostas 69, 1071, Abstract 1884, 1993.