- Email: [email protected]
Antagonism of the platelet 5HT2 receptor in the presence of thrombolysis
FLSEVIFR SC‘IENCF IRFLANI)
International Journal of Cardiology 43 (1994) I I-20
Antagonism of the platelet 5HT2 receptor in the presence of thrombolysis Philip R. Belcher *, Angela J. Drake-Holland, Academic
Medicine.
Charing
Cross & Westminster
Medical
School.
Mark I.M. Noble
369 Fulham
Road, London
SW10 9NH.
UK
(Received 24 June 1993; revision accepted 15 September 1993)
Abstract Thrombolysis of coronary arterial thrombi is often accompanied by occlusion due to continued platelet thrombosis. We modified the Felts’ model of intracoronary thrombosis (critical stenosis with endothelial damage) to produce up
to 35min occlusions of the circumflex coronary artery in seven open-chest anaesthetised dogs. Administration of recombinant plasminogen activator (rtPA 200 rg - kg-’ bolus plus 1 mg. kg-’ ah-’ infusion), without heparin or aspirin, always produced effective thrombolysis. Fibrinogen decreased (P < 0.05). thrombin time increased (P < 0.025). bleeding time trebled (P < 0.025) and initial arterial patency was achieved. Addition of a selective 5HTz antagonist, increased coronary blood flow (P < 0.01) and reduced rethrombosis rate (P < 0.025) but did not affect coagulation or bleeding. The time the vessel spent occluded was significantly decreased (P < 0.01) and correlated with the fibrinogen level (r = 0.97, P < O.Ol), thereby implying the presence of fibrin within the thrombus. After rt-PA was withdrawn, bleeding time and fibrinogen level normalised within 30 and 60 min, respectively, but full coronary patency was maintained. Thus, when rt-PA alone had produced full thrombolysis, 5HTz antagonism prevented intracocornary thrombosis without additional bleeding complications. Key
words:
Platelet; Coronary artery; Serotonin; Thrombosis;
1. Introduction Thrombolysis following myocardial infarction reduces short- and long-term mortality [9,20,25]. However, although thrombolytic drugs achieve initial patency, they neither remove the original stimulus to thrombotic occlusion, nor do they affect the development of platelet-rich thrombus [22,26]; indeed, marked platelet activation has been noted after streptokinase administration [ 141. * Corresponding author 0167-5273/94/$07.00 SSDf
0
0167-5273(93)01898-S
1994
Elsevier Science Ireland
Thrombolysis
It is also suggested that plasmin generation increases the release of thrombin [lo], which itself can be more thrombogenic than the original injury [2]. Because of these factors, rethrombosis can affect a significant percentage of cases [23.37,40] despite concomitant administration of heparin and aspirin. Aspirin does not improve arterial patency after thrombolysis [34]. The benefits of heparin remain to be proved [25]. It is therefore clear that adjunctive treatment is required to prevent rethrombosis (31. The importance of serotonin (5HT) in platelet
Ltd. All rights reserved
12
aggregation was investigated by De Clerck [ 121;he demonstrated that 5HT amplified the aggregatory response of the platelet to ADP, adrenaline and exposed collagen via the platelet 5HTz receptor. To assess the importance of the serotonin (5HTz) receptor in the pathogenesis of acute coronary syndromes, many studies have been undertaken using canine models of coronary artery disease, most of which have featured fixed stenoses overlying endothelial damage [ 1,&l 1,32,41]. In addition to prevention of platelet thrombosis, a probable dispersive effect on intracoronary thrombus has now been shown [4]. The degree of stenosis is also of importance; Benedict & Todd [6] have demonstrated that when the degree of stenosis reached 85%, a selective 5HT, antagonist was more effective than aspirin in the prevention of platelet thrombosis in an animal model. In none of the studies did increased bleeding occur. We postulated that serotonin 5HT2 antagonism was likely to be the ideal adjuvant therapy in coronary arterial thrombotic occlusion. The present study was undertaken to test this hypothesis. We intended to test the hypothesis only and therefore did not wish to rely on one pharmacological tool which might have possessed an unknown idiosyncratic action. To demonstrate the principle involved, we therefore selected three pharmacological agents, all possessing pure 5HT2 antagonist effects, used in random order. Recombinant tissue type plasminogen activator (rt-PA) was chosen as the thrombolytic agent because of the absence of allergic phenomena, and its short half-life (5-6 min). Neither pre-treatment with any anti-platelet agent nor heparinisation was used. 2. Materials and methods 2. I. Flow studies
Eight beagles were anaesthetised with sodium pentobarbitone 20 mg. kg-’ and intubated with a cuffed endotracheal tube. They were ventilated using a Bleese-Manley volume cycled ventilator with 50% N,O/O_, mixture; anaesthesia was also maintained with regular additional sodium pentobarbitone (3 mg. kg-‘) at 30-min intervals. The gas flow and ventilator were adjusted to keep arterial pH, POZ and PCO? in the normal physio-
logical range, and the end-tidal CO? was monitored. Venous access was via the right cephalic vein, and both femoral veins to avoid contamination for drug infusion and blood sampling. The right femoral artery was cannulated and systemic blood pressure measured through a fluid tilled line connected to a Statham P23Db transducer. A standard limb lead electrocardiogram was displayed continuously. The heart in seven dogs was approached by left thoracotomy through the fifth interspace. (One other anaesthetised. but not operated dog, was studied only for the haematological effects of the drugs.) The pericardium was opened and sutured to the wound edges to provide a cradle without displacing the heart. The circumflex coronary artery was exposed in the atrioventricular groove by apical reflexion of its accompanying vein. About 2 cm of the artery which was free of large branches could always be isolated. A 1.5 or 2.0 mm electromagnetic flow probe (Skalar Medical; Delft, Netherlands), according to arterial size, was placed on the artery proximally to measure mean and phasic circumflex coronary artery blood flow. These measurements and mean blood pressure were recorded continuously on a Gould electrostatic ES2000 paper recorder. Just distal to the flow probe, the artery was crushed with a haemostat to produce endothelial damage and then surrounded by a 4.0-mm long split polythene tube. This cylindrical occluder had a 1.O-mm internal diameter and was reclosed by an encircling ligature. A further ligature was passed loosely around the distalmost portion of the artery for the purpose of temporary complete occlusion. The occluder provides a 70-85% stenosis as previously described [41]. Fine adjustments were made by insertion of a tapered nylon fly fishing line within the lumen of the occluder. For these experiments, the stenosis was adjusted so that the thrombus built up to occlude the vessel completely. Usually the vessel was allowed to remain occluded until it embolised spontaneously (Fig. 1). Periodically, if this had not occurred, the vessel was shaken to remove the embolus manually; it was considered important to do this in order to check that the vessel had not been inadvertently mechanically obstructed completely.
P.R. Belcher et al. /Int.
J. Cardiol. 43 (1994)
13
I I-20
This manoeuvre was also performed in the control period to generate several measurable slopes of decline in coronary blood flow. This is a measure of the rate of thrombus growth [4,28,31] and has been demonstrated to be directly proportional to the rate at which platelets accumulate in the arterial lumen [30]. To eliminate the possibility of zero drift in the electromagnetic flow probes, any flow registering less than or equal to 5 ml.min-’ was regarded as zero flow. Because the artery was prepared in order to produce occlusion of the vessel by platelet thrombosis, we also elected to look at the proportion of time that the vessel spent occluded, which was compared at various stages of the investigation. 2.2. Coagulation variables The coagulation variables were measured using an SCL coagulation system (Instrumentation Laboratories, USA). The coefficients of variation for prothrombin, thrombin and partial thromboplastin times are less than 2%, and less than 6% for the fibrinogen estimation. Buccal mucosal bleeding time [27] was assesssed using a Simplate II device (Organon, Teknika; Belgium). Bleeding time measurements which were equal to or greater than 8 min were recorded as 8 min. 2.3. Experimental protocol
The experiment was a repeated measures design with the following stages: 1. Control period with establishment of platelet thrombosis with flow decreasing to, and remaining at, zero. Measurements of haematological variables including bleeding time and blood coagulation factors. 2. Bolus injection of 200 pg.kg-’ rt-PA, followed by a continuous intravenous infusion at 1 mg.kg-‘*h-l for 1 h. 3. (a) Haematological study after 20 min of infusion. (b) Haematological investigation after 40 min of infusion. 4. After 45 min of rt-PA infusion, administration of a bolus dose of 5HT2 antagonist, until flow improved. 5. Fifteen minutes later, cessation of rt-PA
infusion with observation of coronary blood flow. (If full patency was not maintained after this, one further dose of a 5HT, antagonist was administered.) 6. Haematological study at 30 min after cessation of rt-PA infusion. 7. Observation of coronary blood flow and haematological study at 60 min after cessation of rt-PA infusion. recommended human The manufacturer’s dosage schedule for rt-PA (Actilyse; Boehringer, Ingelheim) is a total dose of 1.4 mg - kg-’ administered over 3 h. Ten percent of the total dose is administered as a bolus, 50% is then infused over the first hour and the remainder over the subsequent 2 h. It was therefore decided, for the purposes of this study, to give a 200 pg. kg-’ bolus and then to infuse at 1 mg. kg-’ . h-‘. Forty-five minutes was chosen as the time to record during the effect of rt-PA alone, as this has been shown to be sufficient time to allow full thrombolysis to take effect; such effectiveness was confirmed haematologically in each study. The selective 5HT, antagonists used in the study are (as far as is known) pharmacologically identical and are known potent selective antagonists at 5-hydroxytryptamine receptors of the 5HT2 subtype, with negligable affinity to the other three subtypes of the serotonin receptor. These agents were MDL 11,939: (a-phenyl- 1-(2phenyl-ethyl)4-piperidine methanol), and in two animals, LY53,857 (6-methyl-l-(l-methylethyl)ergoline-8-carboxylic acid 2-hydroxy- 1-methylpropylester-(Z)-2-butenedioate (1:)). or ICI 170,809 (2(2-dimethylamino-2-methylpropylthio)-3-phenylquinoline). Data are presented as means f 1 standard deviation. Statistical analysis of group data was by distribution-free methods. Linear regression was also employed where appropriate. P-values of less than 0.05 were considered significant. 3. Results 3.1. Circumflex coronary artery blood jlow The sequence of events occurring in a typical experiment is shown in Fig. 1. After the adminis-
14
P.R. B&her
et (11./ Int. J. Curdiol. 43 (1994)
II-20
40 Mean circumflex coronary blood flov (ml/mid
3O
2O
d-PA infusion
+5HTz antagonist pso.0 1
Fig. I. This shows the course of an experiment in one dog. Circumflex coronary artery blood flow was measured at one minute intervals throughout. Initially the flow was unstable and decreased to zero as the vessel thtombosed. Intravenous injection of rt-PA bolus (200 &kg) plus infusion at 1.5 mg. kg-‘. h-‘, did not improve flow; 0.8 mg .kg-’ MDL 1I 939 restored flow which was maintained even after cessation of the r&PA infusion.
tration of the bolus and 45 minutes infusion of rtPA, no significant difference in circumflex coronary blood flow was seen, as is shown in Table 1. However, administration of a 5HT2 antagonist whilst the rt-PA was still being infused resulted in a highly significant increase in blood flow (Fig. 2). The proportion of time that the flow in the artery fell to 5 ml .min-’ or less was also highly significantly improved (Fig. 3). This improvement was maintained even after the withdrawal of rt-PA (Table 1). Further administration of 5HT2 antagonist to those animals which still showed signs of
Fig. 2. This shows the universal and highly significant rise in circumflex blood flow which was observed following addition of a selective 5HT, antagonist, after approximately 45 min of r&PA infusion. P < 0.01 (Wilcoxon).
intracoronary platelet thrombosis (i.e. continued reductions in coronary flow after 60 min of rt-PA infusion) resulted in persistent stable flow in the arteries (Table 1). 3.2. Index of thrombosis formation The rate at which coronary blood flow declined (ml.min-*) in the artery under investigation is shown in Table 1. As expected this index of the rate of platelet thrombus growth was not affected by rt-PA administration. Addition of a 5HT2 antagonist significantly reduced the rate of thrombus growth in six dogs with abolition of platelet thrombosis in four of seven animals. Withdrawal of rt-PA was accompanied by an increase in the
Table I Rate of rethrombosis (ml emin-*) and percentage proportion of time vessel was occluded: effects of rt-PA infusion f 5HT2 receptor antagonist Variable (mean f SD.)
Stage of investigation Control
2,3 r&PA alone
4 rt-PA + 5HT2 antagonist
SP rt-PA Off No further SHT? antagonist
6.9 f 2.5 II.2 f 6.5 43 f I9
7.2 zt 2.9 6.2 l 3.7 64 f 24
2.51 f 3.1* 17.0 f 8.1* 25 f l9**
2.7 zt 3.0 15.8 f 8.5 I2 f 21
1
Rethrombosis rate (ml-mitt-*) Mean flow (ml - minm2) Percentage time of or near occlusion
*P C 0.025 vs. preceding stage of investigation. **P < 0.01 vs. preceding stage of investigation.
P.R. Belcher et al. /IN. J. Cardiol. 43 (1994)
100
Proportion of time that vessel vas at or near occlusion (X)
II-20
I5
-
s 80 _
6o ]
h
40
*O-
\
08
01
rt-PA infusion
+5HT antagonist (pcO.01)
Fig. 3. The proportion of time that the blood flow in the coronary artery was zero or less than 5 ml. min-’ is shown on the ordinate. The addition of a selective SHT, antagonist universally resulted in a signiticant improvement in this situation. P < 0.01 (Wilcoxon).
thrombosis rate in two animals, one of which had had platelet thrombosis abolished earlier by the initial dose of 5HT2 antagonist (P = NS). There were two other dogs which still had evidence of platelet thrombosis at this stage; this was completely abolished by further administration of a 5HTz antagonist in all four animals.
Table 2 Template
bleeding
Measurement
point
Bleeding
time (min)
Thrombin
time (s)
Prothrombin Partial
time and coagulation
time (s)
thromboplastin
Fibrinogen
(9. I-‘)
time (s)
studies:
effects of rt-PA
,'
,I’
;
1.2
1.0'
1.4
1.6
Fibrinmgen (g/L)
Fig. 4. This graph shows the highly linear relation between plasma fibrinogen level and the proportion of time the vessel spent at or near occlusion (flow less than 5 mlimin). These values were measured after one hour of rt-PA infusion and 15 min after administration of a selective 5HTz antagonist; (+) mean point of regression; dotted lines. 95% confidence belts.
3.3. Coagulation factors Fibrinogen levels (mean laboratory value for dogs is 1.9 f 0.7 g-1-l) declined steadily during rt-PA infusion; this difference was significant after 20 min and persisted until withdrawal of rt-PA. One hour after stopping rt-PA infusion, the fibrinogen values had returned to levels which were slightly lower, but not significantly different,
infusion
f serotonin,
receptor
antagonist
I
2
3
4
5
6
2.1 ztO.6 14.0 l3 6.0 zto.1 19.0 *4 1.6 zto.3
5.48; zt2.0 21.0** *9 6.0 zto.1 18.0 *4 1.3* *0.2
6.3 *I.8 23.0 *9 6. I zto.1 19.0 *2 1.3 zto.1
6.8 f I.9 26.0 *I3 6.3 zto.5 16.0 l3 1.1; *0.3
2.6*’ zto.7 32.0 l 29 6.1 *O.l 18.0 *I I.2 *0.4
2.6 *0.8 20.0°q l4 6.1 l 0.2 17.0* *2 I .4-l zto.3
The ‘measurement points’ refer to the following stages of the protocol, namely: 1, control; 2, +rt-PA 20 min: 3. +rt-PA 45 min: 4, +rt-PA 60 min + 5HT,; stop rt-PA for 30 min: stop rt-PA for 60 min. +P < 0.05 vs. previous stage. **P < 0.025 vs. previous stage. tP < 0.05 vs. stage 4. $P < 0.01 vs. stage 4. OP < 0.025 vs. control.
16
from control. The fibrinogen values were also compared to the proportion of time the vessel spent occluded. Before rt-PA was administered there was no correlation between fibrinogen level and the proportion of time that the vessel spent occluded. During rt-PA infusion significant correlations existed at 20 and 40 min (r = 0.815, P < 0.05; r = 0.822, P < 0.05, respectively). After administration of the 5HT2 antagonist, the correlation was much stronger (r = 0.97, P < 0.01); this is shown in Fig. 4. Arcsine transformation of the proportion of time that the vessel was occluded, made only trivial differences to the correlation coefficients and none to the significance values. Normal values for dogs, in our laboratory, of prothrombin, thrombin and partial thromboplastin times are 6, 20 and 15 s, respectively. Prothrombin and partial thromboplastin times were unchanged at any stage of the investigation. Thrombin time, however, rose steadily throughout the administration of rt-PA. These thrombin levels declined one hour after the cessation of the rt-PA infusion, and remained significantly higher than control (Table 2). 3.4. Bleeding time The results of the template bleeding times are shown in Table 2. With use of the standardised template bleeding time, control values averaged 2.2 min (our laboratory normal is 2.4 f 0.7 min). Twenty minutes of rt-PA infusion prolonged bleeding time significantly. These values increased further, but not significantly, during further rt-PA administration. Bleeding time was not increased by addition of any selective 5HT2 antagonist. Thirty minutes after withdrawal of rt-PA (six plasma half-lives of the drug), bleeding time had decreased significantly and was comparable to control values. 4. Discussion
The main findings of this study are that when rtPA alone has produced full thrombolysis (indicated by the low fibrinogen levels), the addition of a selective 5HT2 antagonist results in abolition of intracoronary thrombosis. This effect persists despite withdrawal of rt-PA and subsequent normalisation of bleeding times and coagulation.
P.R. B&her
Ed 01. / Inr. J. Cardiol.
43
( 1994)
II-20
Acute platelet thrombus formation in diseased coronary arteries is an important early step in the development of coronary thrombotic occlusion [21]. Platelet-rich thrombi appear at sites of narrowing associated with endothelial denudation. In order to study platelet aggregation in vivo, it was decided to use a canine model in which a critical coronary artery stenosis overlies a segment of damaged endothelium. This model was modified from its original concept [ 151 in order to provide from ten to thirty minutes of occlusion. This model has subsequently been shown to correlate closely with acute ischaemic syndromes in humans, with similar clinical outcomes of which coronary arterial occlusion is prominent [24]. Thus the model is of considerable relevance in the design of new therapeutic strategies to overcome the persistent clinical coronary arterial occlusion problems that occur in these syndromes with present treatment modalities. The present study indicates that selective 5HT2 receptor antagonism is predicted to be a solution to these coronary arterial occlusion problems. A previous study used a possibly less clinically relevant model. Copper coils have been inserted in the unstenosed vessel to produce a fibrin-rich thrombus. Using this model, the effect of pretreatment by 5HT2 and thromboxane A2 antagonists has been studied by Golino et al [19]. Thrombolysis with heparin and rt-PA was only improved by pretreatment with both agents [l9]. This method is of particular relevance when assessing librinolytic agents by themselves. In contrast, the present investigation attempts to examine plateletmediated rethrombosis in conditions of severe stenosis and endothelial damage. It is possible that dogs are hypercoagulable due to the normally low thrombin, partial thromboplastin, and prothrombin times. We have previously measured these variables and established normal ranges for the dog, which are lower than human values. This potential hypercoagulability, however, may be offset by the low fibrinogen levels seen in these animals, leading to the slightly lower bleeding time. Despite these species differences, there was clear evidence of increased bleeding time and decreased fibrinogen levels during rt-PA administration (see Table 2). It can be considered that the dose regimen rt-PA
P.R. Belcher et al. /Int.
J. Cardiol.
43 (1994)
I I-20
used in the present study would be more than adequate to ensure thrombolysis [ 181; furthermore, the increase seen in thrombin time indicates that the functional integrity of circulating fibrinogen has been altered [38]. The correlation between fibrinogen level and proportion of time the vessel spent occluded, during the total period of rt-PA infusion, was weak until the selective 5HT2 antagonist was given (Fig. 3). This is compatible with the demonstrated continuance of platelet thrombosis during rt-PA infusion. The strong correlation seen may indicate that the platelet thrombus produced in the modification of Folts’ model used in the present study is enriched with fibrin; this is also likely to follow from the known presence of fibrin in the Folts model thrombus [ 161. Platelet thrombosis in the unmodified model appears independent of heparin and the coronary blood flow reductions are repeatable and cyclical; the thrombi have been shown to contain mainly platelets, with some fibrin, red and white blood cells [16]. There is no evidence that the effects are due to spasm [ 161. The vessel wall adopts a convoluted position within a rigid plastic ring so that changes in vessel wall tone have no effect on the resistance to blood flow across the stenosis. This was checked in separate experiments in which a distal pressure line was used to study the flow and pressure gradient across the stenosis [5]. In addition, flow is not affected by the direct application of the smooth muscle relaxant papaverine to the artery within the rigid ring. rt-PA initiated flow when administered at the time of arterial occlusion, but the rate of thrombus formation in this preparation was unaffected. As rt-PA is a fibrinolytic, its actions on a platelet-rich thrombus would be expected to be slight. However, periods of occlusion lasting up to 35 min were observed. Flow values of less than 5 ml.min-’ were considered as zero flow (to allow for zero drift) when cycling was not occurring. This may have been due either to no flow at all, or to a very low flow, presumably through an almost occluded artery; this would evoke high shear stress rates as shear stress increases with reduction of intraluminal diameter. This would stimulate further thrombosis and occlusion of the vessel [30]. We have previously measured stenosis (proximal) and downstream (distal) resistance in the Folts
17
model during thrombus growth [5]. It was found, during cyclic flow reductions, that downstream resistance remained constant while stenosis resistance increased and thereby decreased flow. Because of the long periods of occlusion, it might be expected that fibrin would form an element of the thrombus. That rt-PA failed to maintain coronary artery patency under these circumstances supports the necessity to supplement thrombolysis with an effective antiplatelet agent. Clinically, aspirin is used routinely as an adjunct to thrombolysis, but does not reduce time to reperfusion or have persistent effects in a dog thrombolysis model [44]. Aspirin does not improve coronary artery patency following thrombolytic therapy for myocardial infarction [34]. Aspirin increases the template bleeding time following thrombolysis [17], but its anti-aggregatory effects on platelets are easily overcome by adrenaline in physiological concentrations [28,36]. The relative ineffectiveness of aspirin has prompted investigation of other platelet receptors. 5HT, antagonism was very effective in establishing and maintaining flow and might therefore be considered in preference to other approaches. The pattern of platelet thrombosis is unaffected by heparin [ 161. However, specific thrombin inhibition by argatroban has been demonstrated to be effective in decreasing time to reperfusion in an experimental model, but it considerably increases the activated partial thromboplastin and bleeding times. Hirudin has not been shown to be effective at preventing post-thrombotic reocclusion despite prolongation of the bleeding time [39]. The combination of thromboxane synthetase inhibition and receptor blockade (R68 070, (ridogrel)), or receptor blockade alone by SQ 29 548 abolishes platelet thrombosis but is associated with bleeding [35]. The synthetic prostaglandin analogue, iloprost, has been shown to improve patency after thrombolysis, but is associated with coagulation derangement and a fall in haemoglobin concentration [29]. Blockade of the platelet GPIIb/IIIa receptor causes bleeding despite successful lysis of platelet-rich coronary thrombus resistant to plasminogen activator [44]. The platelet contains large amounts of dissolved serotonin and the dense granules have a very high concentration of serotonin. Platelet release thus
18
results in very high local levels of serotonin (11 which may form the basis of the amplification of the platelet activation and aggregation reactions. The effect of serotonin antagonism on coronary arterial tone was not measured in this study. Changes in coronary vascular tone are not responsible for the cycling of flow because they cannot account for the sudden extremely rapid restorations of flow that occur and which are caused by embolisation of thrombus. The preceding flow reduction period is thus caused by thrombus growth. Restoration of flow by 5HT2 antagonism is thus caused by removal of thrombus. In a previous study of the Folts model [5], downstream resistance was constant during thrombosis; platelet thrombosis was shown to increase this resistance value slightly, although significantly, compared to undamaged arteries in which thrombosis was not occurring. This difference in downstream resistance, with or without a proximal thrombus is so small that its possible removal by a selective 5HT2 antagonist would not affect the results reported in the present study; they are accounted for by removal of proximal resistance, i.e., thrombus. Thus, proximal resistance is the flow limiting resistance during the period of declining flow (cyclic flow reduction). The main influence on blood flow during cyclic flow reductions was the presence or absence of thrombosis at the stenosis, Therefore, flow depended on stenosis resistance. Flow rates after 5HT, antagonist administration never exceeded the highest flows seen beforehand in the preparation, confirming a negligible effect of 5HT, antagonism on coronary vascular resistance. Despite the classification of serotonin as a weak agonist of platelet aggregation in vitro, there is evidence that its effects are of great importance in amplication of platelet activation and aggregation [12]. LY 53 857 in combination with a thromboxane A, antagonist, can improve efficacy of rt-PA thrombolysis and abolish rethrombosis in the copper coil canine model [ 191. Three selective 5HT, receptor antagonists, LY 53 857, ICI 170 809 and MDL 11 939 had no obvious coagulation defects noted on preliminary investigations [7,13,43]. This lack of effect on coagulation factors or bleeding time is confirmed by the present investigation. The significance of these results is
P.R. B&her
PI ul. / Int. J. Cur&l.
43
f 1994 I 11-X
enhanced by two findings: first, bleeding was not worsened by addition of the drug to an already deranged clotting system; and second, after normalization of bleeding time and coagulation variables with return to near normal fibrinogen levels, reocclusion did not occur; instead there was persistent coronary artery blood flow with no evidence of recurrent thrombosis. Our experiences, both in the laboratory [4,41] and clinically [33] led us to think that the use of selective serotonin 5HT2 antagonists, by themselves, gives great promise for adjunctive treatment of intracoronary thrombosis, without worsening concomitant coagulation defects. In contrast to aspirin, heparin, thromboxane inhibitors and most other antiplatelet agents, selective 5HT, receptor antagonists are not known to affect haemostasis. The effectiveness of 5HT2 antagonism in the presence of heparin and aspirin has not been assessed but cannot be any less than that of 5HTz antagonism alone. 5HT2 antagonists have been shown to be synergistic with aspirin [42], but heparin is an independent anticoagulant with no direct anti-platelet effects [16]. 5. Acknowledgements rt-PA was generously donated by Dr D.M. Humphreys of Boehringer Ingelheim. 6. References Ashton JH, Benedict CR, Fitzgerald C et al. Serotonin as a mediator of cyclic flow variations in stenosed canine coronary arteries. Circulation 1986; 73: 572-578. Badimon L, Lassila R, Badimon J, Vallabhajosula B, Chesebro JH, Fuster V. Residual thrombus is more thrombogenic than severely damaged vessel wall [abstract]. Circulation 1988; 78 Suppl II: 11-l 19. Becker RC. Gore JM. Adjuvant antiplatelet strategies in coronary thrombolysis [editorial]. Circulation 1991; 83: 1115-1117. Belcher PR, Drake-Holland AJ. Hynd JW, Noble MIM. Dispersion of coronary artery thrombi by antagonism of platelet serotonin receptor in the dog. Cardiovasc Res 1992: 26: 292-296. Belcher PR, Vergroesen I, Hynd JW. Coronary arterial pressure-flow relation distal to a critical stenosis: influence of thrombosis in the anaesthetized dog. J Physiol 1993; 467: 316P. Benedict CR, Todd GE. Effect of aspirin and serotonin receptor (5HT,) blockade on coronary occlusion by thrombus formation [abstract]. Circulation 1991; 84 Suppl II: 11-79.
P.R. Belcher et al. /Int.
7
8
9
10
I1
12 13
14
15
16
17
18
19
20
21
J. Cardiol.
43 (1994)
II-20
Blackburn TP, Haworth SJ, Jessup CL, Morton PB, Williams C. ICI 170809, a selective 5-hydroxytryptamine antagonist, inhibits human platelet aggregation in vitro and ex vivo. In: Saxena PR, Wallis DI. Wouters W, Bevan P, editors. Cardiovascular pharmacology of 5-hydroxytryptamine. Dordrecht: Kluwer Academic, 1990; 459-463. Bush LR. Effects of the serotonin antagonists, cyproheptadine, ketanserin and mianserin. on cyclic flow reductions in stenosed canine coronary arteries. J Pharmacol Exptl Therap 1987; 240: 674-682. Chesebro JH, Knatterud G, Roberts R, Circulation et al. Thrombolysis in myocardial infarction (TIMI) trial, Phase I: a comparison between intravenous tissue plasminogen activator and intravenous streptokinase. Circulation 1987; 76: 42-54. Chesebro JH, Zoldhelyi P, Fuster V. Pathogenesis of thrombosis in unstable angina. Am J Cardiol 1991; 68: 2B-10B. Cox B, McAulitTe SJG, Noble MIM, Snow HM. Inhibition of coronary arterial platelet thrombosis in the anaesthetised dog by antagonism of 5HTz receptors. J Physiol 1991; 438: 54P. De Clerck F, Herman AC. 5-Hydroxytryptamine and platelet aggregation. Fed Proc 1983; 42: 228-232. Dudley MW, Wiech NL, Miller FP et al. Pharmacological effects of MDL I 1.939: a selective, centrally acting antagonist of 5-HT, receptors. Drug Develop Res 1988; 13: 29-43. Fitzgerald DJ, Catella F, Roy L, Fitzgerald GA. Marked platelet activation in vivo after intravenous streptokinase in patients with acute myocardial infarction. Circulation 1988; 77: 142-150. Felts JD, Crowell EB, Rowe GC. Platelet aggregation in partially obstructed vessels and its elimination with aspirin. Circulation 1976; 54: 365-370. Folts JD, Gallagher K, Rowe GG. Blood flow reductions in stenosed canine coronary arteries: vasospasm or platelet aggregation. Circulation 1982; 65: 248-255. Garabedian HD, Gold HK, Leinbach RC et al. Bleeding time prolongation and bleeding during infusion of recombinant tissue-type plasminogen activator in dogs: potentiation by aspirin and reversal with aprotinin. J Am Co11 Cardiol 1991; 17: 1213-1222. Gold HK. Fallon JT. Yashuda T et al. Coronary thrombolysis with recombinant human tissue-type plasminogen activator. Circulation 1984; 70: 700-707. Golino P, Ashton JH, McNatt J et al. Simultaneous administration of thromboxane A2- and serotonin S2receptor antagonists markedly enhances thrombolysis and prevents or delays reocclusion after tissue-type plasminogen activator in a canine model of coronary thrombosis. Circulation 1989; 79: 91 l-919. Gruppo Italian0 per lo Studio della Sopravvivenza nell’lnfarcto Miocardico (GISSI). Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Lancet 1986; i: 397-402. Haerem JW. Platelet aggregates in intramyocardial vessels
19
22
23
24
25
26
27
28
29
30
31
32
33
34
of patients dying suddenly and unexpectedly of coronary artery disease. Atherosclerosis 1972; 15: 199-213. Haskel J, Adams SP, Feigen LP et al. Prevention of platelet-rich thrombi in canine femoral arteries with a novel peptide antagonist of platelet glycoprotein Ilbillla receptors. Circulation 1989; 80: 1775-1782. Hsia J, Hamilton WP, Kleiman N, Roberts R, Chaitman BR, Ross AM. A comparison between heparin and lowdose aspirin as adjunctive therapy with tissue plasminogen activator for acute myocardial infarction. Heparinaspirin reperfusion trial (HART) investigators. N Engl J Med 1990; 323: 1433-1437. Ikeda H, Koga Y, Kuwano K et al. Cyclic flow variations in a conscious dog model of coronary artery stenoses and endothelial injury correlate with acute ischemic heart disease syndromes in humans. J Am Coll Cardiol 1993: 21: 1008-1017. ISIS-3: a randomised comparison of streptokinase vs. tissue plasminogen activator vs. anistreplase and of aspirin plus heparin vs. aspirin alone among 41,299 cases of suspected acute myocardial infarction. Lancet 1992; 339: 753-770. Jang I-K, Gold HK, Leinbach RC, Rivera AG, Fallon JT, Collen D. Lysis of resistant thrombus by t-PA combined with platelet Ilb/Illa receptor blocking synthetic pentapeptide (G4120) and selective thrombin inhibitor, argatroban. Circulation 1991: 84 Suppl II: B-599. Jergens AE, Turrentine MA, Kraus KH, Johnson GS. Buccal mucosa bleeding times of healthy dogs and of dogs in various pathologic states, including thrombocytopenia. uremia, and von Willebrand’s disease. Am J Vet Res 1987; 48: 1337-1342. Keller JW. Felts JD. Combined inhibitory effects of aspirin and ethanol on adrenaline exacerbation of acute platelet thrombus formation in stenosed canine coronary arteries. Cardiovasc Res 1990; 24: 191-197. Kerins DM, Roy L. Kunitada S et al. Pharmacokinetics of tissue-type plasminogen activator during acute myocardial infarction in men. Effect of a prostacyclin analogue. Circulation 1992; 85: 526-532. Maalej N. Folts JD. Aspirin treatment does not protect against acute platelet thrombosis in high shear coronary artery stenosis: quantitative assessment. Thromb Haemostasis 1993; 69: 588A. McAuliffe SJG. Snow HM, Smith CTT. Noble MIM. Interaction between the effects of 5-hydroxytryptamine and adrenaline on the growth of platelet thrombi in the coronary artery of the anaesthetized dog. Br J Pharmacol 1993; 109: 405-410. Morishima Y, Tanaka T. Watanabe K, Igarishi T. Yasuoka M, Shibano T. Prevention by DV-7028, a selective 5-HT, receptor antagonist, of the formation of coronary thrombi in dogs. Cardiovasc Res 1991; 25: 727-730. Noble MIM. Drake-Holland AJ. The involvement of serotonin in the formation of thrombi at critical coronary artery stenoses in humans. Coronary Artery Dis 1990; I: 675-679. Norris RM. White HD. Cross DB. et al. Aspirin does not
20
35
36
37
38
39
P.R. Bekher
improve early arterial patency after streptokinase treatment for acute myocardial infarction. Br Heart J 1993; 69: 492-495. Patron0 C. Biosynthesis and pharmacological modulation of thromboxane in humans. Circulation 1990; 81 Suppl I: 1-12-I-15. Rao GHR, Escolar G, White JG. Epinephrine reverses the inhibitory influence of aspirin on platelet-vessel wall interactions. Thrombosis Res 1986; 44: 65-74. Schaer DH, Ross AM, Wasserman AG. Reinfarction, recurrent angina and reocclusion after thrombolytic therapy. Circulation 1987; 76 Suppl II: H-57-11-62. Shafer KE, Santoro SA, Sobel BE, Jaffe AS. Monitoring activity of Iibrinolytic agents: a therapeutic challenge. Am J Med 1984; 76: 879-886. Sitko GR, Ramjit DR. Stabilito II, Lehman D, Lynch JJ, Vlasuk GP. Conjunctive enhancement of enzymatic thrombolysis and prevention of thrombotic reocclusion with the selective factor Xa inhibitor, tick anticoagulant peptide. Circulation 1992; 85: 805-8 15.
40
41
42 43
44
CI trl. / Inr. J. Curdid.
43
f 1994)
11-20
Tiefenbrunn AJ, Sobel BE. The impact of coronary thrombolysis in myocardial infarction. Fibrinolysis 1989; 3: I-15. Torr S, Noble MIM, Folts JD. Inhibition of acute platelet thrombosis formation in stenosed canine coronary arteries by specific 5HT, receptor antagonist ritanserin. Cardiovasc Res 1990; 24: 465-470. Willerson JT. Serotonin and thrombotic complications. J Cardiovasc Pharmacol 1991; 17 Suppl 5: Sl3-S20. Yao S-K. Rosolowsky M, Anderson HV et al. Combined thromboxane A2 synthetase inhibition and receptor blockade are effective in preventing spontaneous and epinephrine-induced canine coronary cyclic flow variations. J Am Co11 Cardiol 1990; 16: 705-713. Yasuda T, Gold HK, Yaoita H et al. Comparative effects of aspirin, a synthetic thrombin inhibitor and a monoclonal antiplatelet glycoprotein IIb/IIIa antibody on coronary artery reperfusion, reocclusion and bleeding with recombinant tissue-type plasminogen activator in a canine preparation. J Am Co11 Cardiol 1990: 16: 714-722.
International Journal of Cardiology 43 (1994) I I-20
Antagonism of the platelet 5HT2 receptor in the presence of thrombolysis Philip R. Belcher *, Angela J. Drake-Holland, Academic
Medicine.
Charing
Cross & Westminster
Medical
School.
Mark I.M. Noble
369 Fulham
Road, London
SW10 9NH.
UK
(Received 24 June 1993; revision accepted 15 September 1993)
Abstract Thrombolysis of coronary arterial thrombi is often accompanied by occlusion due to continued platelet thrombosis. We modified the Felts’ model of intracoronary thrombosis (critical stenosis with endothelial damage) to produce up
to 35min occlusions of the circumflex coronary artery in seven open-chest anaesthetised dogs. Administration of recombinant plasminogen activator (rtPA 200 rg - kg-’ bolus plus 1 mg. kg-’ ah-’ infusion), without heparin or aspirin, always produced effective thrombolysis. Fibrinogen decreased (P < 0.05). thrombin time increased (P < 0.025). bleeding time trebled (P < 0.025) and initial arterial patency was achieved. Addition of a selective 5HTz antagonist, increased coronary blood flow (P < 0.01) and reduced rethrombosis rate (P < 0.025) but did not affect coagulation or bleeding. The time the vessel spent occluded was significantly decreased (P < 0.01) and correlated with the fibrinogen level (r = 0.97, P < O.Ol), thereby implying the presence of fibrin within the thrombus. After rt-PA was withdrawn, bleeding time and fibrinogen level normalised within 30 and 60 min, respectively, but full coronary patency was maintained. Thus, when rt-PA alone had produced full thrombolysis, 5HTz antagonism prevented intracocornary thrombosis without additional bleeding complications. Key
words:
Platelet; Coronary artery; Serotonin; Thrombosis;
1. Introduction Thrombolysis following myocardial infarction reduces short- and long-term mortality [9,20,25]. However, although thrombolytic drugs achieve initial patency, they neither remove the original stimulus to thrombotic occlusion, nor do they affect the development of platelet-rich thrombus [22,26]; indeed, marked platelet activation has been noted after streptokinase administration [ 141. * Corresponding author 0167-5273/94/$07.00 SSDf
0
0167-5273(93)01898-S
1994
Elsevier Science Ireland
Thrombolysis
It is also suggested that plasmin generation increases the release of thrombin [lo], which itself can be more thrombogenic than the original injury [2]. Because of these factors, rethrombosis can affect a significant percentage of cases [23.37,40] despite concomitant administration of heparin and aspirin. Aspirin does not improve arterial patency after thrombolysis [34]. The benefits of heparin remain to be proved [25]. It is therefore clear that adjunctive treatment is required to prevent rethrombosis (31. The importance of serotonin (5HT) in platelet
Ltd. All rights reserved
12
aggregation was investigated by De Clerck [ 121;he demonstrated that 5HT amplified the aggregatory response of the platelet to ADP, adrenaline and exposed collagen via the platelet 5HTz receptor. To assess the importance of the serotonin (5HTz) receptor in the pathogenesis of acute coronary syndromes, many studies have been undertaken using canine models of coronary artery disease, most of which have featured fixed stenoses overlying endothelial damage [ 1,&l 1,32,41]. In addition to prevention of platelet thrombosis, a probable dispersive effect on intracoronary thrombus has now been shown [4]. The degree of stenosis is also of importance; Benedict & Todd [6] have demonstrated that when the degree of stenosis reached 85%, a selective 5HT, antagonist was more effective than aspirin in the prevention of platelet thrombosis in an animal model. In none of the studies did increased bleeding occur. We postulated that serotonin 5HT2 antagonism was likely to be the ideal adjuvant therapy in coronary arterial thrombotic occlusion. The present study was undertaken to test this hypothesis. We intended to test the hypothesis only and therefore did not wish to rely on one pharmacological tool which might have possessed an unknown idiosyncratic action. To demonstrate the principle involved, we therefore selected three pharmacological agents, all possessing pure 5HT2 antagonist effects, used in random order. Recombinant tissue type plasminogen activator (rt-PA) was chosen as the thrombolytic agent because of the absence of allergic phenomena, and its short half-life (5-6 min). Neither pre-treatment with any anti-platelet agent nor heparinisation was used. 2. Materials and methods 2. I. Flow studies
Eight beagles were anaesthetised with sodium pentobarbitone 20 mg. kg-’ and intubated with a cuffed endotracheal tube. They were ventilated using a Bleese-Manley volume cycled ventilator with 50% N,O/O_, mixture; anaesthesia was also maintained with regular additional sodium pentobarbitone (3 mg. kg-‘) at 30-min intervals. The gas flow and ventilator were adjusted to keep arterial pH, POZ and PCO? in the normal physio-
logical range, and the end-tidal CO? was monitored. Venous access was via the right cephalic vein, and both femoral veins to avoid contamination for drug infusion and blood sampling. The right femoral artery was cannulated and systemic blood pressure measured through a fluid tilled line connected to a Statham P23Db transducer. A standard limb lead electrocardiogram was displayed continuously. The heart in seven dogs was approached by left thoracotomy through the fifth interspace. (One other anaesthetised. but not operated dog, was studied only for the haematological effects of the drugs.) The pericardium was opened and sutured to the wound edges to provide a cradle without displacing the heart. The circumflex coronary artery was exposed in the atrioventricular groove by apical reflexion of its accompanying vein. About 2 cm of the artery which was free of large branches could always be isolated. A 1.5 or 2.0 mm electromagnetic flow probe (Skalar Medical; Delft, Netherlands), according to arterial size, was placed on the artery proximally to measure mean and phasic circumflex coronary artery blood flow. These measurements and mean blood pressure were recorded continuously on a Gould electrostatic ES2000 paper recorder. Just distal to the flow probe, the artery was crushed with a haemostat to produce endothelial damage and then surrounded by a 4.0-mm long split polythene tube. This cylindrical occluder had a 1.O-mm internal diameter and was reclosed by an encircling ligature. A further ligature was passed loosely around the distalmost portion of the artery for the purpose of temporary complete occlusion. The occluder provides a 70-85% stenosis as previously described [41]. Fine adjustments were made by insertion of a tapered nylon fly fishing line within the lumen of the occluder. For these experiments, the stenosis was adjusted so that the thrombus built up to occlude the vessel completely. Usually the vessel was allowed to remain occluded until it embolised spontaneously (Fig. 1). Periodically, if this had not occurred, the vessel was shaken to remove the embolus manually; it was considered important to do this in order to check that the vessel had not been inadvertently mechanically obstructed completely.
P.R. Belcher et al. /Int.
J. Cardiol. 43 (1994)
13
I I-20
This manoeuvre was also performed in the control period to generate several measurable slopes of decline in coronary blood flow. This is a measure of the rate of thrombus growth [4,28,31] and has been demonstrated to be directly proportional to the rate at which platelets accumulate in the arterial lumen [30]. To eliminate the possibility of zero drift in the electromagnetic flow probes, any flow registering less than or equal to 5 ml.min-’ was regarded as zero flow. Because the artery was prepared in order to produce occlusion of the vessel by platelet thrombosis, we also elected to look at the proportion of time that the vessel spent occluded, which was compared at various stages of the investigation. 2.2. Coagulation variables The coagulation variables were measured using an SCL coagulation system (Instrumentation Laboratories, USA). The coefficients of variation for prothrombin, thrombin and partial thromboplastin times are less than 2%, and less than 6% for the fibrinogen estimation. Buccal mucosal bleeding time [27] was assesssed using a Simplate II device (Organon, Teknika; Belgium). Bleeding time measurements which were equal to or greater than 8 min were recorded as 8 min. 2.3. Experimental protocol
The experiment was a repeated measures design with the following stages: 1. Control period with establishment of platelet thrombosis with flow decreasing to, and remaining at, zero. Measurements of haematological variables including bleeding time and blood coagulation factors. 2. Bolus injection of 200 pg.kg-’ rt-PA, followed by a continuous intravenous infusion at 1 mg.kg-‘*h-l for 1 h. 3. (a) Haematological study after 20 min of infusion. (b) Haematological investigation after 40 min of infusion. 4. After 45 min of rt-PA infusion, administration of a bolus dose of 5HT2 antagonist, until flow improved. 5. Fifteen minutes later, cessation of rt-PA
infusion with observation of coronary blood flow. (If full patency was not maintained after this, one further dose of a 5HT, antagonist was administered.) 6. Haematological study at 30 min after cessation of rt-PA infusion. 7. Observation of coronary blood flow and haematological study at 60 min after cessation of rt-PA infusion. recommended human The manufacturer’s dosage schedule for rt-PA (Actilyse; Boehringer, Ingelheim) is a total dose of 1.4 mg - kg-’ administered over 3 h. Ten percent of the total dose is administered as a bolus, 50% is then infused over the first hour and the remainder over the subsequent 2 h. It was therefore decided, for the purposes of this study, to give a 200 pg. kg-’ bolus and then to infuse at 1 mg. kg-’ . h-‘. Forty-five minutes was chosen as the time to record during the effect of rt-PA alone, as this has been shown to be sufficient time to allow full thrombolysis to take effect; such effectiveness was confirmed haematologically in each study. The selective 5HT, antagonists used in the study are (as far as is known) pharmacologically identical and are known potent selective antagonists at 5-hydroxytryptamine receptors of the 5HT2 subtype, with negligable affinity to the other three subtypes of the serotonin receptor. These agents were MDL 11,939: (a-phenyl- 1-(2phenyl-ethyl)4-piperidine methanol), and in two animals, LY53,857 (6-methyl-l-(l-methylethyl)ergoline-8-carboxylic acid 2-hydroxy- 1-methylpropylester-(Z)-2-butenedioate (1:)). or ICI 170,809 (2(2-dimethylamino-2-methylpropylthio)-3-phenylquinoline). Data are presented as means f 1 standard deviation. Statistical analysis of group data was by distribution-free methods. Linear regression was also employed where appropriate. P-values of less than 0.05 were considered significant. 3. Results 3.1. Circumflex coronary artery blood jlow The sequence of events occurring in a typical experiment is shown in Fig. 1. After the adminis-
14
P.R. B&her
et (11./ Int. J. Curdiol. 43 (1994)
II-20
40 Mean circumflex coronary blood flov (ml/mid
3O
2O
d-PA infusion
+5HTz antagonist pso.0 1
Fig. I. This shows the course of an experiment in one dog. Circumflex coronary artery blood flow was measured at one minute intervals throughout. Initially the flow was unstable and decreased to zero as the vessel thtombosed. Intravenous injection of rt-PA bolus (200 &kg) plus infusion at 1.5 mg. kg-‘. h-‘, did not improve flow; 0.8 mg .kg-’ MDL 1I 939 restored flow which was maintained even after cessation of the r&PA infusion.
tration of the bolus and 45 minutes infusion of rtPA, no significant difference in circumflex coronary blood flow was seen, as is shown in Table 1. However, administration of a 5HT2 antagonist whilst the rt-PA was still being infused resulted in a highly significant increase in blood flow (Fig. 2). The proportion of time that the flow in the artery fell to 5 ml .min-’ or less was also highly significantly improved (Fig. 3). This improvement was maintained even after the withdrawal of rt-PA (Table 1). Further administration of 5HT2 antagonist to those animals which still showed signs of
Fig. 2. This shows the universal and highly significant rise in circumflex blood flow which was observed following addition of a selective 5HT, antagonist, after approximately 45 min of r&PA infusion. P < 0.01 (Wilcoxon).
intracoronary platelet thrombosis (i.e. continued reductions in coronary flow after 60 min of rt-PA infusion) resulted in persistent stable flow in the arteries (Table 1). 3.2. Index of thrombosis formation The rate at which coronary blood flow declined (ml.min-*) in the artery under investigation is shown in Table 1. As expected this index of the rate of platelet thrombus growth was not affected by rt-PA administration. Addition of a 5HT2 antagonist significantly reduced the rate of thrombus growth in six dogs with abolition of platelet thrombosis in four of seven animals. Withdrawal of rt-PA was accompanied by an increase in the
Table I Rate of rethrombosis (ml emin-*) and percentage proportion of time vessel was occluded: effects of rt-PA infusion f 5HT2 receptor antagonist Variable (mean f SD.)
Stage of investigation Control
2,3 r&PA alone
4 rt-PA + 5HT2 antagonist
SP rt-PA Off No further SHT? antagonist
6.9 f 2.5 II.2 f 6.5 43 f I9
7.2 zt 2.9 6.2 l 3.7 64 f 24
2.51 f 3.1* 17.0 f 8.1* 25 f l9**
2.7 zt 3.0 15.8 f 8.5 I2 f 21
1
Rethrombosis rate (ml-mitt-*) Mean flow (ml - minm2) Percentage time of or near occlusion
*P C 0.025 vs. preceding stage of investigation. **P < 0.01 vs. preceding stage of investigation.
P.R. Belcher et al. /IN. J. Cardiol. 43 (1994)
100
Proportion of time that vessel vas at or near occlusion (X)
II-20
I5
-
s 80 _
6o ]
h
40
*O-
\
08
01
rt-PA infusion
+5HT antagonist (pcO.01)
Fig. 3. The proportion of time that the blood flow in the coronary artery was zero or less than 5 ml. min-’ is shown on the ordinate. The addition of a selective SHT, antagonist universally resulted in a signiticant improvement in this situation. P < 0.01 (Wilcoxon).
thrombosis rate in two animals, one of which had had platelet thrombosis abolished earlier by the initial dose of 5HT2 antagonist (P = NS). There were two other dogs which still had evidence of platelet thrombosis at this stage; this was completely abolished by further administration of a 5HTz antagonist in all four animals.
Table 2 Template
bleeding
Measurement
point
Bleeding
time (min)
Thrombin
time (s)
Prothrombin Partial
time and coagulation
time (s)
thromboplastin
Fibrinogen
(9. I-‘)
time (s)
studies:
effects of rt-PA
,'
,I’
;
1.2
1.0'
1.4
1.6
Fibrinmgen (g/L)
Fig. 4. This graph shows the highly linear relation between plasma fibrinogen level and the proportion of time the vessel spent at or near occlusion (flow less than 5 mlimin). These values were measured after one hour of rt-PA infusion and 15 min after administration of a selective 5HTz antagonist; (+) mean point of regression; dotted lines. 95% confidence belts.
3.3. Coagulation factors Fibrinogen levels (mean laboratory value for dogs is 1.9 f 0.7 g-1-l) declined steadily during rt-PA infusion; this difference was significant after 20 min and persisted until withdrawal of rt-PA. One hour after stopping rt-PA infusion, the fibrinogen values had returned to levels which were slightly lower, but not significantly different,
infusion
f serotonin,
receptor
antagonist
I
2
3
4
5
6
2.1 ztO.6 14.0 l3 6.0 zto.1 19.0 *4 1.6 zto.3
5.48; zt2.0 21.0** *9 6.0 zto.1 18.0 *4 1.3* *0.2
6.3 *I.8 23.0 *9 6. I zto.1 19.0 *2 1.3 zto.1
6.8 f I.9 26.0 *I3 6.3 zto.5 16.0 l3 1.1; *0.3
2.6*’ zto.7 32.0 l 29 6.1 *O.l 18.0 *I I.2 *0.4
2.6 *0.8 20.0°q l4 6.1 l 0.2 17.0* *2 I .4-l zto.3
The ‘measurement points’ refer to the following stages of the protocol, namely: 1, control; 2, +rt-PA 20 min: 3. +rt-PA 45 min: 4, +rt-PA 60 min + 5HT,; stop rt-PA for 30 min: stop rt-PA for 60 min. +P < 0.05 vs. previous stage. **P < 0.025 vs. previous stage. tP < 0.05 vs. stage 4. $P < 0.01 vs. stage 4. OP < 0.025 vs. control.
16
from control. The fibrinogen values were also compared to the proportion of time the vessel spent occluded. Before rt-PA was administered there was no correlation between fibrinogen level and the proportion of time that the vessel spent occluded. During rt-PA infusion significant correlations existed at 20 and 40 min (r = 0.815, P < 0.05; r = 0.822, P < 0.05, respectively). After administration of the 5HT2 antagonist, the correlation was much stronger (r = 0.97, P < 0.01); this is shown in Fig. 4. Arcsine transformation of the proportion of time that the vessel was occluded, made only trivial differences to the correlation coefficients and none to the significance values. Normal values for dogs, in our laboratory, of prothrombin, thrombin and partial thromboplastin times are 6, 20 and 15 s, respectively. Prothrombin and partial thromboplastin times were unchanged at any stage of the investigation. Thrombin time, however, rose steadily throughout the administration of rt-PA. These thrombin levels declined one hour after the cessation of the rt-PA infusion, and remained significantly higher than control (Table 2). 3.4. Bleeding time The results of the template bleeding times are shown in Table 2. With use of the standardised template bleeding time, control values averaged 2.2 min (our laboratory normal is 2.4 f 0.7 min). Twenty minutes of rt-PA infusion prolonged bleeding time significantly. These values increased further, but not significantly, during further rt-PA administration. Bleeding time was not increased by addition of any selective 5HT2 antagonist. Thirty minutes after withdrawal of rt-PA (six plasma half-lives of the drug), bleeding time had decreased significantly and was comparable to control values. 4. Discussion
The main findings of this study are that when rtPA alone has produced full thrombolysis (indicated by the low fibrinogen levels), the addition of a selective 5HT2 antagonist results in abolition of intracoronary thrombosis. This effect persists despite withdrawal of rt-PA and subsequent normalisation of bleeding times and coagulation.
P.R. B&her
Ed 01. / Inr. J. Cardiol.
43
( 1994)
II-20
Acute platelet thrombus formation in diseased coronary arteries is an important early step in the development of coronary thrombotic occlusion [21]. Platelet-rich thrombi appear at sites of narrowing associated with endothelial denudation. In order to study platelet aggregation in vivo, it was decided to use a canine model in which a critical coronary artery stenosis overlies a segment of damaged endothelium. This model was modified from its original concept [ 151 in order to provide from ten to thirty minutes of occlusion. This model has subsequently been shown to correlate closely with acute ischaemic syndromes in humans, with similar clinical outcomes of which coronary arterial occlusion is prominent [24]. Thus the model is of considerable relevance in the design of new therapeutic strategies to overcome the persistent clinical coronary arterial occlusion problems that occur in these syndromes with present treatment modalities. The present study indicates that selective 5HT2 receptor antagonism is predicted to be a solution to these coronary arterial occlusion problems. A previous study used a possibly less clinically relevant model. Copper coils have been inserted in the unstenosed vessel to produce a fibrin-rich thrombus. Using this model, the effect of pretreatment by 5HT2 and thromboxane A2 antagonists has been studied by Golino et al [19]. Thrombolysis with heparin and rt-PA was only improved by pretreatment with both agents [l9]. This method is of particular relevance when assessing librinolytic agents by themselves. In contrast, the present investigation attempts to examine plateletmediated rethrombosis in conditions of severe stenosis and endothelial damage. It is possible that dogs are hypercoagulable due to the normally low thrombin, partial thromboplastin, and prothrombin times. We have previously measured these variables and established normal ranges for the dog, which are lower than human values. This potential hypercoagulability, however, may be offset by the low fibrinogen levels seen in these animals, leading to the slightly lower bleeding time. Despite these species differences, there was clear evidence of increased bleeding time and decreased fibrinogen levels during rt-PA administration (see Table 2). It can be considered that the dose regimen rt-PA
P.R. Belcher et al. /Int.
J. Cardiol.
43 (1994)
I I-20
used in the present study would be more than adequate to ensure thrombolysis [ 181; furthermore, the increase seen in thrombin time indicates that the functional integrity of circulating fibrinogen has been altered [38]. The correlation between fibrinogen level and proportion of time the vessel spent occluded, during the total period of rt-PA infusion, was weak until the selective 5HT2 antagonist was given (Fig. 3). This is compatible with the demonstrated continuance of platelet thrombosis during rt-PA infusion. The strong correlation seen may indicate that the platelet thrombus produced in the modification of Folts’ model used in the present study is enriched with fibrin; this is also likely to follow from the known presence of fibrin in the Folts model thrombus [ 161. Platelet thrombosis in the unmodified model appears independent of heparin and the coronary blood flow reductions are repeatable and cyclical; the thrombi have been shown to contain mainly platelets, with some fibrin, red and white blood cells [16]. There is no evidence that the effects are due to spasm [ 161. The vessel wall adopts a convoluted position within a rigid plastic ring so that changes in vessel wall tone have no effect on the resistance to blood flow across the stenosis. This was checked in separate experiments in which a distal pressure line was used to study the flow and pressure gradient across the stenosis [5]. In addition, flow is not affected by the direct application of the smooth muscle relaxant papaverine to the artery within the rigid ring. rt-PA initiated flow when administered at the time of arterial occlusion, but the rate of thrombus formation in this preparation was unaffected. As rt-PA is a fibrinolytic, its actions on a platelet-rich thrombus would be expected to be slight. However, periods of occlusion lasting up to 35 min were observed. Flow values of less than 5 ml.min-’ were considered as zero flow (to allow for zero drift) when cycling was not occurring. This may have been due either to no flow at all, or to a very low flow, presumably through an almost occluded artery; this would evoke high shear stress rates as shear stress increases with reduction of intraluminal diameter. This would stimulate further thrombosis and occlusion of the vessel [30]. We have previously measured stenosis (proximal) and downstream (distal) resistance in the Folts
17
model during thrombus growth [5]. It was found, during cyclic flow reductions, that downstream resistance remained constant while stenosis resistance increased and thereby decreased flow. Because of the long periods of occlusion, it might be expected that fibrin would form an element of the thrombus. That rt-PA failed to maintain coronary artery patency under these circumstances supports the necessity to supplement thrombolysis with an effective antiplatelet agent. Clinically, aspirin is used routinely as an adjunct to thrombolysis, but does not reduce time to reperfusion or have persistent effects in a dog thrombolysis model [44]. Aspirin does not improve coronary artery patency following thrombolytic therapy for myocardial infarction [34]. Aspirin increases the template bleeding time following thrombolysis [17], but its anti-aggregatory effects on platelets are easily overcome by adrenaline in physiological concentrations [28,36]. The relative ineffectiveness of aspirin has prompted investigation of other platelet receptors. 5HT, antagonism was very effective in establishing and maintaining flow and might therefore be considered in preference to other approaches. The pattern of platelet thrombosis is unaffected by heparin [ 161. However, specific thrombin inhibition by argatroban has been demonstrated to be effective in decreasing time to reperfusion in an experimental model, but it considerably increases the activated partial thromboplastin and bleeding times. Hirudin has not been shown to be effective at preventing post-thrombotic reocclusion despite prolongation of the bleeding time [39]. The combination of thromboxane synthetase inhibition and receptor blockade (R68 070, (ridogrel)), or receptor blockade alone by SQ 29 548 abolishes platelet thrombosis but is associated with bleeding [35]. The synthetic prostaglandin analogue, iloprost, has been shown to improve patency after thrombolysis, but is associated with coagulation derangement and a fall in haemoglobin concentration [29]. Blockade of the platelet GPIIb/IIIa receptor causes bleeding despite successful lysis of platelet-rich coronary thrombus resistant to plasminogen activator [44]. The platelet contains large amounts of dissolved serotonin and the dense granules have a very high concentration of serotonin. Platelet release thus
18
results in very high local levels of serotonin (11 which may form the basis of the amplification of the platelet activation and aggregation reactions. The effect of serotonin antagonism on coronary arterial tone was not measured in this study. Changes in coronary vascular tone are not responsible for the cycling of flow because they cannot account for the sudden extremely rapid restorations of flow that occur and which are caused by embolisation of thrombus. The preceding flow reduction period is thus caused by thrombus growth. Restoration of flow by 5HT2 antagonism is thus caused by removal of thrombus. In a previous study of the Folts model [5], downstream resistance was constant during thrombosis; platelet thrombosis was shown to increase this resistance value slightly, although significantly, compared to undamaged arteries in which thrombosis was not occurring. This difference in downstream resistance, with or without a proximal thrombus is so small that its possible removal by a selective 5HT2 antagonist would not affect the results reported in the present study; they are accounted for by removal of proximal resistance, i.e., thrombus. Thus, proximal resistance is the flow limiting resistance during the period of declining flow (cyclic flow reduction). The main influence on blood flow during cyclic flow reductions was the presence or absence of thrombosis at the stenosis, Therefore, flow depended on stenosis resistance. Flow rates after 5HT, antagonist administration never exceeded the highest flows seen beforehand in the preparation, confirming a negligible effect of 5HT, antagonism on coronary vascular resistance. Despite the classification of serotonin as a weak agonist of platelet aggregation in vitro, there is evidence that its effects are of great importance in amplication of platelet activation and aggregation [12]. LY 53 857 in combination with a thromboxane A, antagonist, can improve efficacy of rt-PA thrombolysis and abolish rethrombosis in the copper coil canine model [ 191. Three selective 5HT, receptor antagonists, LY 53 857, ICI 170 809 and MDL 11 939 had no obvious coagulation defects noted on preliminary investigations [7,13,43]. This lack of effect on coagulation factors or bleeding time is confirmed by the present investigation. The significance of these results is
P.R. B&her
PI ul. / Int. J. Cur&l.
43
f 1994 I 11-X
enhanced by two findings: first, bleeding was not worsened by addition of the drug to an already deranged clotting system; and second, after normalization of bleeding time and coagulation variables with return to near normal fibrinogen levels, reocclusion did not occur; instead there was persistent coronary artery blood flow with no evidence of recurrent thrombosis. Our experiences, both in the laboratory [4,41] and clinically [33] led us to think that the use of selective serotonin 5HT2 antagonists, by themselves, gives great promise for adjunctive treatment of intracoronary thrombosis, without worsening concomitant coagulation defects. In contrast to aspirin, heparin, thromboxane inhibitors and most other antiplatelet agents, selective 5HT, receptor antagonists are not known to affect haemostasis. The effectiveness of 5HT2 antagonism in the presence of heparin and aspirin has not been assessed but cannot be any less than that of 5HTz antagonism alone. 5HT2 antagonists have been shown to be synergistic with aspirin [42], but heparin is an independent anticoagulant with no direct anti-platelet effects [16]. 5. Acknowledgements rt-PA was generously donated by Dr D.M. Humphreys of Boehringer Ingelheim. 6. References Ashton JH, Benedict CR, Fitzgerald C et al. Serotonin as a mediator of cyclic flow variations in stenosed canine coronary arteries. Circulation 1986; 73: 572-578. Badimon L, Lassila R, Badimon J, Vallabhajosula B, Chesebro JH, Fuster V. Residual thrombus is more thrombogenic than severely damaged vessel wall [abstract]. Circulation 1988; 78 Suppl II: 11-l 19. Becker RC. Gore JM. Adjuvant antiplatelet strategies in coronary thrombolysis [editorial]. Circulation 1991; 83: 1115-1117. Belcher PR, Drake-Holland AJ. Hynd JW, Noble MIM. Dispersion of coronary artery thrombi by antagonism of platelet serotonin receptor in the dog. Cardiovasc Res 1992: 26: 292-296. Belcher PR, Vergroesen I, Hynd JW. Coronary arterial pressure-flow relation distal to a critical stenosis: influence of thrombosis in the anaesthetized dog. J Physiol 1993; 467: 316P. Benedict CR, Todd GE. Effect of aspirin and serotonin receptor (5HT,) blockade on coronary occlusion by thrombus formation [abstract]. Circulation 1991; 84 Suppl II: 11-79.
P.R. Belcher et al. /Int.
7
8
9
10
I1
12 13
14
15
16
17
18
19
20
21
J. Cardiol.
43 (1994)
II-20
Blackburn TP, Haworth SJ, Jessup CL, Morton PB, Williams C. ICI 170809, a selective 5-hydroxytryptamine antagonist, inhibits human platelet aggregation in vitro and ex vivo. In: Saxena PR, Wallis DI. Wouters W, Bevan P, editors. Cardiovascular pharmacology of 5-hydroxytryptamine. Dordrecht: Kluwer Academic, 1990; 459-463. Bush LR. Effects of the serotonin antagonists, cyproheptadine, ketanserin and mianserin. on cyclic flow reductions in stenosed canine coronary arteries. J Pharmacol Exptl Therap 1987; 240: 674-682. Chesebro JH, Knatterud G, Roberts R, Circulation et al. Thrombolysis in myocardial infarction (TIMI) trial, Phase I: a comparison between intravenous tissue plasminogen activator and intravenous streptokinase. Circulation 1987; 76: 42-54. Chesebro JH, Zoldhelyi P, Fuster V. Pathogenesis of thrombosis in unstable angina. Am J Cardiol 1991; 68: 2B-10B. Cox B, McAulitTe SJG, Noble MIM, Snow HM. Inhibition of coronary arterial platelet thrombosis in the anaesthetised dog by antagonism of 5HTz receptors. J Physiol 1991; 438: 54P. De Clerck F, Herman AC. 5-Hydroxytryptamine and platelet aggregation. Fed Proc 1983; 42: 228-232. Dudley MW, Wiech NL, Miller FP et al. Pharmacological effects of MDL I 1.939: a selective, centrally acting antagonist of 5-HT, receptors. Drug Develop Res 1988; 13: 29-43. Fitzgerald DJ, Catella F, Roy L, Fitzgerald GA. Marked platelet activation in vivo after intravenous streptokinase in patients with acute myocardial infarction. Circulation 1988; 77: 142-150. Felts JD, Crowell EB, Rowe GC. Platelet aggregation in partially obstructed vessels and its elimination with aspirin. Circulation 1976; 54: 365-370. Folts JD, Gallagher K, Rowe GG. Blood flow reductions in stenosed canine coronary arteries: vasospasm or platelet aggregation. Circulation 1982; 65: 248-255. Garabedian HD, Gold HK, Leinbach RC et al. Bleeding time prolongation and bleeding during infusion of recombinant tissue-type plasminogen activator in dogs: potentiation by aspirin and reversal with aprotinin. J Am Co11 Cardiol 1991; 17: 1213-1222. Gold HK. Fallon JT. Yashuda T et al. Coronary thrombolysis with recombinant human tissue-type plasminogen activator. Circulation 1984; 70: 700-707. Golino P, Ashton JH, McNatt J et al. Simultaneous administration of thromboxane A2- and serotonin S2receptor antagonists markedly enhances thrombolysis and prevents or delays reocclusion after tissue-type plasminogen activator in a canine model of coronary thrombosis. Circulation 1989; 79: 91 l-919. Gruppo Italian0 per lo Studio della Sopravvivenza nell’lnfarcto Miocardico (GISSI). Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Lancet 1986; i: 397-402. Haerem JW. Platelet aggregates in intramyocardial vessels
19
22
23
24
25
26
27
28
29
30
31
32
33
34
of patients dying suddenly and unexpectedly of coronary artery disease. Atherosclerosis 1972; 15: 199-213. Haskel J, Adams SP, Feigen LP et al. Prevention of platelet-rich thrombi in canine femoral arteries with a novel peptide antagonist of platelet glycoprotein Ilbillla receptors. Circulation 1989; 80: 1775-1782. Hsia J, Hamilton WP, Kleiman N, Roberts R, Chaitman BR, Ross AM. A comparison between heparin and lowdose aspirin as adjunctive therapy with tissue plasminogen activator for acute myocardial infarction. Heparinaspirin reperfusion trial (HART) investigators. N Engl J Med 1990; 323: 1433-1437. Ikeda H, Koga Y, Kuwano K et al. Cyclic flow variations in a conscious dog model of coronary artery stenoses and endothelial injury correlate with acute ischemic heart disease syndromes in humans. J Am Coll Cardiol 1993: 21: 1008-1017. ISIS-3: a randomised comparison of streptokinase vs. tissue plasminogen activator vs. anistreplase and of aspirin plus heparin vs. aspirin alone among 41,299 cases of suspected acute myocardial infarction. Lancet 1992; 339: 753-770. Jang I-K, Gold HK, Leinbach RC, Rivera AG, Fallon JT, Collen D. Lysis of resistant thrombus by t-PA combined with platelet Ilb/Illa receptor blocking synthetic pentapeptide (G4120) and selective thrombin inhibitor, argatroban. Circulation 1991: 84 Suppl II: B-599. Jergens AE, Turrentine MA, Kraus KH, Johnson GS. Buccal mucosa bleeding times of healthy dogs and of dogs in various pathologic states, including thrombocytopenia. uremia, and von Willebrand’s disease. Am J Vet Res 1987; 48: 1337-1342. Keller JW. Felts JD. Combined inhibitory effects of aspirin and ethanol on adrenaline exacerbation of acute platelet thrombus formation in stenosed canine coronary arteries. Cardiovasc Res 1990; 24: 191-197. Kerins DM, Roy L. Kunitada S et al. Pharmacokinetics of tissue-type plasminogen activator during acute myocardial infarction in men. Effect of a prostacyclin analogue. Circulation 1992; 85: 526-532. Maalej N. Folts JD. Aspirin treatment does not protect against acute platelet thrombosis in high shear coronary artery stenosis: quantitative assessment. Thromb Haemostasis 1993; 69: 588A. McAuliffe SJG. Snow HM, Smith CTT. Noble MIM. Interaction between the effects of 5-hydroxytryptamine and adrenaline on the growth of platelet thrombi in the coronary artery of the anaesthetized dog. Br J Pharmacol 1993; 109: 405-410. Morishima Y, Tanaka T. Watanabe K, Igarishi T. Yasuoka M, Shibano T. Prevention by DV-7028, a selective 5-HT, receptor antagonist, of the formation of coronary thrombi in dogs. Cardiovasc Res 1991; 25: 727-730. Noble MIM. Drake-Holland AJ. The involvement of serotonin in the formation of thrombi at critical coronary artery stenoses in humans. Coronary Artery Dis 1990; I: 675-679. Norris RM. White HD. Cross DB. et al. Aspirin does not
20
35
36
37
38
39
P.R. Bekher
improve early arterial patency after streptokinase treatment for acute myocardial infarction. Br Heart J 1993; 69: 492-495. Patron0 C. Biosynthesis and pharmacological modulation of thromboxane in humans. Circulation 1990; 81 Suppl I: 1-12-I-15. Rao GHR, Escolar G, White JG. Epinephrine reverses the inhibitory influence of aspirin on platelet-vessel wall interactions. Thrombosis Res 1986; 44: 65-74. Schaer DH, Ross AM, Wasserman AG. Reinfarction, recurrent angina and reocclusion after thrombolytic therapy. Circulation 1987; 76 Suppl II: H-57-11-62. Shafer KE, Santoro SA, Sobel BE, Jaffe AS. Monitoring activity of Iibrinolytic agents: a therapeutic challenge. Am J Med 1984; 76: 879-886. Sitko GR, Ramjit DR. Stabilito II, Lehman D, Lynch JJ, Vlasuk GP. Conjunctive enhancement of enzymatic thrombolysis and prevention of thrombotic reocclusion with the selective factor Xa inhibitor, tick anticoagulant peptide. Circulation 1992; 85: 805-8 15.
40
41
42 43
44
CI trl. / Inr. J. Curdid.
43
f 1994)
11-20
Tiefenbrunn AJ, Sobel BE. The impact of coronary thrombolysis in myocardial infarction. Fibrinolysis 1989; 3: I-15. Torr S, Noble MIM, Folts JD. Inhibition of acute platelet thrombosis formation in stenosed canine coronary arteries by specific 5HT, receptor antagonist ritanserin. Cardiovasc Res 1990; 24: 465-470. Willerson JT. Serotonin and thrombotic complications. J Cardiovasc Pharmacol 1991; 17 Suppl 5: Sl3-S20. Yao S-K. Rosolowsky M, Anderson HV et al. Combined thromboxane A2 synthetase inhibition and receptor blockade are effective in preventing spontaneous and epinephrine-induced canine coronary cyclic flow variations. J Am Co11 Cardiol 1990; 16: 705-713. Yasuda T, Gold HK, Yaoita H et al. Comparative effects of aspirin, a synthetic thrombin inhibitor and a monoclonal antiplatelet glycoprotein IIb/IIIa antibody on coronary artery reperfusion, reocclusion and bleeding with recombinant tissue-type plasminogen activator in a canine preparation. J Am Co11 Cardiol 1990: 16: 714-722.