THROMBOSIS
RESEARCH
61; 501-514,199l
0049-3848/91 $3.00 + .OOPrinted in the USA. Copyright
(c) 1991 Pergamon
Press pk. All rights reserved.
CHANGES IN PLATELET FUNCTION AND REACTIVITY INDUCED BY QUININE IN RELATION TO QUININE (DRUG) INDUCED IMMUNE THROMBOCYTOPENIA
J.M. Connellan, S. Deacon & P.J. Thurlow Department of Haematology, Austin Hospital, Studley Road, Heidelberg, Victoria, 3084, Australia (Received
7.9.1990;
accepted
in revised form 13.12.1990
by Editor H.H. Salem)
ABSTRACT Quinine, a drug known to induce immune mediated thrombocytopenia, has been postulated to mediate binding of drug dependent antibodies to a range of platelet membrane glycoproteins. Quinine may not act solely as a hapten however, as we have shown that it inhibits platelet aggregation (ex vivo and in vitro) and release and modifies the ability of activated platelets to bind the adhesive proteins fibrinogen and fibronectin in a dose dependent fashion. Studies on the effect of quinine on the binding of monoclonal antibodies HuPlml (Gpllla) FMC25 (GplX) and AN51 (Gplb) to platelets shows a selective reduction in AN51 binding. In addition quinine induced platelet antibodies from thrombocytopenic patients, in the presence of quinine, have been shown to inhibit binding of these monoclonal antibodies to platelets to varying degrees. These observations suggest that quinine causes widespread but specific conformational changes in platelet membrane antigens resulting in the production of quinine which may expose neoantigens induced antibodies.
INTRODUCTION Quinine and quinidine (cinchona alkaloid stereoisomers) have been demonstrated to induce anti-platelet antibodies resulting in severe thrombocytopenia in susceptible patients (1,2). Patients with quinine-induced thrombocytopenia have
Key words: Platelet antibody, quinine, thrombocytopenia, monoclonal antibodies to membrane glycoproteins. 501
fibrinogen,
fibronectin,
502
QUININE INDUCED THROMBOCYTOPENIA
Vol61, Nos. 516
been shown to possess a heterogeneous population of antibodies, with specificities for platelet glycoprotein (Gp) Ilb, Illa, lb, IX, V (3,4,5,6), which require the presence of the sensitizing drug for maximum binding of antibody to platelet. The initial studies (7) used to demonstrate the existence of quinine and quinidine antibodies relied upon platelet aggregometry where a platelet proaggregant effect was induced by adding patients’ serum together with quinine or quinidine to normal platelet rich plasma. However the effects of quinine or quinidine on the platelet membrane and the possible role this may play in the production of antibodies has been largely ignored. It has been assumed that quinine or quinidine act as haptens in the binding of antibodies to the platelet, by complexing with another moiety presumably either a plasma protein, or a protein on the platelet surface (8,9). However, no quinine-binding plasma factor has been demonstrated and Christie and co-workers (10,ll) have shown that quinine binds only weakly to platelet membranes. It is possible that quinine has an effect on the production of antibodies unrelated to its previously proposed role as a hapten. Quinine may have a direct effect in vitro and in vivo on platelet membrane proteins in which platelet structure and possibly function are effected. These changes could occur as a consequence of perturbation of membrane glycoproteins with concomitant exposure of cryptic antigens. In a similar fashion EDTA has been postulated to induce the exposure, in some patients, of a platelet cryptic antigen. Onder et al (12) demonstrated EDTA associated pseudothrombocytopenia is an h vitro phenomenon, and antiplatelet antibodies only react with platelets in the presence of EDTA (13). Similarly quinine mediated platelet antibodies may develop after exposure of a putative cryptic antigen in platelets caused by the presence of quinine. Our study was undertaken to investigate the effect of quinine (in vitro and ex W) on platelet structure and function as this may indicate the mechanism involvedin production of quinine induced antiplatelet antibodies. Our investigations demonstrate that quinine has marked effects on platelet function and also appears to directly alter membrane antigenicity thus supporting the concept that conformational alterations in platelet membrane antigens in the presence of quinine are important in the induction of quinine induced thrombocytopenia.
MATERIAL AND METHODS Quinine HCI was either modified Tyrode’s 2mM MgCL2) or Hepes KCI) or veronal buffered
purchased from ICN Pharmaceuticals Inc. and dissolved in buffer pH 7.4 (0.15 M NaCI, 2.6 mM KCI, 12 mM NaHC03, buffered saline pH 7.4 (10 mM Hepes, 0.15 M NaCI, 2.6 mM saline pH 7.4 as required.
Platelet rich plasma (PRP) was prepared from blood from either Platelet aaareaation. healthy donors who had not taken any medication in the previous two weeks or healthy volunteers who had ingested 600 mg of quinine bisulphate daily for four days. PRP, anticoagulated with sodium citrate (3.1%), was stored in a stoppered tube at 37” until used. Platelet aggregation studies were performed [l] prior to the ingestion of quinine, [2] whilst the subjects were ingesting quinine and [3] five days after cessation of quinine ingestion. The effect of quinine in vitro on platelets from healthy donors was studied by incubating 5Oyl of quinine (varying concentrations) or Hepes buffered saline with 400 ~1of PRP prior to the addition of 50 f_rlof agonist. Platelet aggregation was followed by using a Chronolog Lumi Aggregometer. Agonists used were (final concentrations): ADP 2, 4, 10 ,uM (Sigma) adrenaline 60 PM (Sigma), collagen 1.6, 8 pg/ml (Hormon-Chemie), arachidonic acid 500 yg/ml
Vol61, Nos. 516
QUININE INDUCED THROMBOCYTOPENIA
503
(Biodata), ristocetin 1 mg/mI (Wellcome) and thrombin 0.2 units/ml (Human plasma, Sigma). 146 serotonin (Amersham S.A. 57 mCi/m mole) was added Platelet release studies. to PRP (20 ~1-1 ).rCi/5 ml) and incubated for 30 minutes at 37”. PRP (400~1) was incubated with varying concentrations of quinine (50~1) prior to the initiation of aggregation with agonist (50~1); aggregation was monitored on a Chronolog Lumi Aggregometer; 250 ,ul of 4.09% EDTA was added to stop aggregation and the mixture centrifuged at 9000 x g for 5 minutes. The supernatant (500~1) was transferred to a scintillation vial together with 500 yl of distilled water and 10 ml of scintillation solution (toluene 2 parts, Triton x-100 1 part; PPO (2,5-diphenyloxazole) 4g/litre) (14) then mixed and counted. A blank (unstimulated, platelet pellet) and total (unstimulated platelets included) tubes were also counted. The percentage release of t4C-5HT was calculated as Test-Blank/Total Blank x 100. Human fibrinogen (Grade L) was obtained from Kabi (Stockholm) Adhesive proteins. and further purified by adsorption on gelatin-Sepharose GB-CL (Pharmacia) in order Fibronectin was purchased from Bethesda to remove fibronectin contamination. Research Laboratories (Maryland USA). Proteins were labelled with 125fodine using the chloramine T Protein labelling. method (modified from Hunter and Greenwood, 15) in brief; fibrinogen (1 mg), fibronectin or IgG (100 t.rg) were incubated with 51.11 of t25lodine 0.5 m Ci(Amersham) and 3 ~1 of chloramine T (lmg/ml); the mixture was incubated for between 2 and 7 minutes at room temperature, then 3yl of sodium metabisulphate (2.4 mg/mI) and 100 ~1 of potassium iodide (16mglml) were added and the mixture was incubated for between 5 and 15 minutes at room temperature. Labelled protein was separated from free iodine by chromatography on a PDlO column (Pharmacia). Gel filtered platelets were prepared from PRP after washing by Gel filtered platelets. centrifugation and resuspended in modified Tyrode’s buffer pH 6.4, 1% BSA. The platelets were separated from contaminating plasma by gel filtration chromatography on Sephadex 2B CL equilibrated with modified Tyrode’s buffer pH 7.4 1% BSA. Fibrinoaen bindina to activated olatelets. Gel filtered platelets (1.4 x 106) suspended in modified Tyrode’s buffer (with 5.5 mM glucose and 0.3% BSA but no MgC12) were preincubated with varying amounts of quinine (0 to 1 .O mM). Activated platelets were prepared by incubating collagen (1.6 yg/mI) with gel filtered platelets at 37” for 3 minutes. 125i fibrinogen was then added to the activated platelets and incubated for a further 30 minutes at 37” and platelet bound 1251 fibrinogen, in the platelet pellet, was measured after centrifugation through phthalate oil (S.G. 1.021). Fibronectin binding to activated olatelets . This was carried out using a similar method to that employed for fibrinogen binding with the exception that thrombin (1 unit/ml) was used to activate platelets and the modified Tyrode’s buffer contained 5.5 mM glucose and 1% BSA. Purification of IaG from plasmas of patients with drua dependent antibodies fDDab). Plasmas were applied to a Protein A Sepharose 66 CL column (Pharmacia), equilibrated with phosphate buffered saline pH 7.4, and incubated for 60 minutes at 4°C. Unbound proteins were eluted with phosphate buffered saline and 0.1 M glycine
504
QUININE INDUCED THROMBOCYTOPENIA
Vol61, Nos. 516
HCI pH 3.0 was used to elute bound IgG which was then dialysed against phosphate buffered saline and stored at -70°C. Purification of monoclonal antibodies (Mab) from ascites or serum. HuPlml, directed against Gpllla, was produced by the authors (16). AN51 directed to Gplb and FMC25 (ascites), directed to GplX, were gifts from Dr. M. Berndt, Westmead Hospital, N.S.W. Mabs were applied to Protein A Sepharose 6B CL column equilibrated with 0.1 M phosphate buffer pH 8.2, and incubated overnight at 4°C; unbound protein was eluted with phosphate buffer. Bound Mabs were eluted with 0.1 M citrate buffer pH 5.8 and dialysed against phosphate buffered saline pH 7.4. The purified Mabs were stored at -70°C at a concentration of 1 mg/ml. Drua Dependent Antibodies bindinq to platelets. Washed platelets were prepared from blood collected into EDTA and resuspended in Tyrode’s buffer pH 7.4. Platelets (1 x 108) were incubated for 30 minutes at 37°C with varying amounts of quinine (O1 .O mM) and DDab plasma or 300 mg of purified IgG (normal plasmas and IgG were used as controls). 1251 Staphylococcal Protein A (SPA) was added and incubated for 30 minutes at 37°C; platelet bound SPA was quantitated after the platelets were pelleted through phthalate oil (Density 1.021) (17). Effect of quinine on the bindinq of Mabs to platelets. Purified Mabs HuPIMI, AN51 and FMC25 were iodinated using the chloramine T method and used in binding experiments to platelets in the presence of varying concentrations of quinine using purified mouse IgG as a control. Washed platelets (1 x 108) resuspended in modified Tyrode’s buffer pH 7.4, 1% BSA were incubated with varying concentrations of quinine (O-l .O mM) and CaC12 (1 mM) following which dilutions of 1251 labelled Mabs were added and incubated for 30 minutes at 37°C. Platelet bound Mab was quantitated after the platelets were pelleted through phthalate oil.(See above) Blockinq of 1251 Mab bindina to platelets by preincubation of olatelets with DDab. Washed platelets (1 x 108) in Tyrode’s buffer pH 7.4, 1% BSA were incubated for 30 minutes at 37°C with plasma or purified IgG (300 r-19)(from patients with DDab or controls) in the presence of l.OmM quinine. Platelets were washed and incubated with 1251 labelled Mab for a further 30 minutes at 37°C. Platelet bound Mab was quantitated after the platelets were pelleted through phthalate oil.
RESULTS PRP, preincubated with Effect of auinine on platelet aaareaation and release. quinine at varying concentrations,was tested with a range of agonists in order to determine a dose response of quinine in platelet aggregation (See Fig 1). At 0.05 mM quinine, which is within the pharmacological range, minimal inhibition was observed only with adrenaline induced aggregation, this level of quinine did not affect aggregation induced by other agonists tested. (ADP (2,4 PM), collagen (1.6, 8.yg/ml), thrombin (0.2U/ml) and ristocetin (lmg/ml) results not shown). At 0.25 mM quinine, adrenaline induced aggregation was largely abolished and slight inhibition was observed in aggregation induced by other agonists. Higher concentrations of quinine (1 .OmM) effectively inhibited aggregation induced by all agonists tested and with 4 PM ADP slight reversible aggregation was demonstrated. Preincubation of PRP with varying amounts of quinine inhibited the release of 14CSerotonin from platelets activated with collagen. Serotonin release, induced by
Vol61, Nos. S/6
505
QUININE INDUCED THROMBOCYTOPENIA
low doses of collagen (1.6 pg/ml), was inhibited by 0.25 mM quinine, whereas 0.5 mM quinine was required to inhibit release induced by higher concentrations of collagen (8 pglml) (Fig. 2). C
B
A Adrenaline I~OJJM 1
Collagen 1Bug/ml)
QmM
ADP
QmM 0, 0.25
0
14!~M) QmM
005
o? O-25
0
12
3
4
0
5
12
3
4
5
(minutes)
Time
FIG. 1 Representative examples of the dose dependent effect of quinine on platelet aggregation induced by weak and strong agonists. PRP was incubated with varying amounts of quinine prior to the addition of agonists (A) adrenaline (60 PM), (B) collagen (8 pg/ml) or (C) ADP (4 PM) 70 I
0 0.0
I 0.2
I 0.4
I 0.6
I 0.8
Quinine (mM) FIG. 2 The effect of quinine on platelet release. 14CSerotonin loaded platelet were incubated with varying concentrations of quinine prior to the addition of collagen (0) = 1.6 pglml (0) = 8 pglml
506
Vol61, Nos. 516
QUININE INDUCED THROMBOCYTOPENIA
Aggregation of PRP initiated by most agonists was reversible when quinine (l.OmM) was added after aggregation had commenced. Aggregation induced by collagen 1.6yg, ADP 2,4 yM, thrombin 0.2 U/ml (data not shown) and adrenaline (60 PM) (Fig 38) were reversed by the addition of quinine. Whilst aggregation induced by collagen 8pg (Fig 3A) and ristocetin lmg/mI (not shown) were not affected by adding 1 mM quinine.
B
A Collagen
Adrenaline
(8Ng/ml)
(60 uM 1
Q (ImM)
/ 3
0
I
1
I
I
1
2
3
4
I
I
5
6
Time
I
I
0 (minutes
1
2
3
1
;
k
)
FIG. 3 Reversal of aggregation by quinine. Quinine (1 .O mM) was added during the course of (A) collagen (8 pg/ml) and (B) adrenaline (60 PM) induced platelet aggregation.
PRP collected from volunteers was studied in triplicate and the ingestion of quinine was found to diminish the aggregatory response to ADP, adrenaline and low dose collagen (1.6 pg/ml). Platelets exposed to quinine in vivo failed to respond to ADP 1pM (Fig 4A) and one out of three samples failed to aggregate when challenged with ADP 2yM whilst platelets from the other two volunteers showed a reduced response. Platelets from all volunteers showed diminished response (average 20% reduction) to ADP 4 I.IM which was a consistent finding. Similarly platelets tested during quinine ingestion, showed a diminished responsiveness to adrenaline, however, there was variation in the extent of aggregation in the individuals tested. Platelets from one volunteer did not respond to adrenaline 1.25 yM,(Fig 48) although aggregation did proceed in the other two volunteers it was accompanied by an extended lag period. All volunteers achieved therapeutic levels of quinine. Some reduction (average 30% decrease) in aggregation occurred after ingestion of quinine when low dose collagen (1.6 pg/mI) was used to stimulate platelets (Fig 4C) and this was overcome by increasing the concentration of collagen to 8 yg/ml (Fig 4D). The platelets responded normally to other agonists, i.e. thrombin, arachidonic acid and A23187. When platelets were tested 5 days after quinine
Vol6 1, Nos. 516
QUININE INDUCED THROMBOCYTOPENIA
507
ingestion was completed, the aggregation response to ADP, adrenaline and collagen returned to normal levels.
A
AOP
IlOpMi
ADRENALINE
B
i 1 25 uM
I
8!.;--”
40
Pre postQuinine
n
Qulnlne lnqestion
1
COLLAGEN
c
(1-61Jg/mli
I
I
0
0
1
2
3
4
5
6
I8 ug/mi I
COLLAGEN
801
Pre + Post 1 Quinine
,/, c
1
2 Trme
3
4
Im~nutes 1
0
1
2 Ttme
,
,
,
3
L
5
6
I mu-&es)
FIG. 4 Representative examples of the effect of quinine ingestion on platelet aggregation measured ex vivo prior to, during and 5 days after completion of ingestion of quinine (600 mg/day) Agonists employed were (A) 1 .O FM ADP, (B) 1.25 PM adrenaline, (C) 1.6 t.rg/ml collagen and (D) 8pg/ml collagen.
Effect of quinine on the bindina of adhesive oroteins to activated platelets. Gel filtered platelets when activated with either collagen or thrombin, without stirring, were induced to bind radiolabelled fibrinogen or fibronectin. Preincubation of platelets with varying amounts of quinine prior to the addition of agonist, resulted in the inhibition of binding of these adhesive proteins. Quinine (0.5 mM) almost completely abolished the binding of fibrinogen stimulated by collagen (1.6 pg/mI) (Fig 5), whereas higher concentrations of quinine (0.8 mM) appear to be necessary to block the binding of fibronectin to thrombin (lU/ml) stimulated platelets (Fig 6). When ImM quinine was added to platelets, which had previously been exposed to low dose collagen and labelled fibrinogen, 73% of the platelet bound fibrinogen was eluted. In contrast the level of bound labelled fibronectin was unaffected by quinine added 30 minutes after activated platelets had been exposed to fibronectin.
508
QUININE INDUCED THROMBOCYTOPENIA
0
i B .E .$
9 0.0
1
I
I
I
I
0.2
0.4
0.6
0.6
1.0
Quinine (mM)
0.0
0.2
0.4
0.6
Quinine (mM)
0.6
1.0
Vol61, Nos. 516
FtG. 5 The effect of quinine on the binding of fibrinogen to platelets. Gel filtered platelets were preincubated with varying amounts of quinine prior to the addition of agonists and labelled adhesive proteins. 1251 fibrinogen was added to collagen (1.6 ,ug/ml) stimulated platelets
FIG. 6 The effect of quinine on the binding of fibronectin to platelets. Gel filtered platelets were preincubated with varying amounts of quinine prior to the addition of agonists and labelled adhesive proteins. 1251 fibronectin was added to thrombin (1 unit/ml) stimulated platelets (0) or resting platelets (e)
The effect of quinine on the binding of Mabs directed to olatelet membrane glycoproteins. Mabs directed to platelet membrane glycoproteins were used to study possible perturbation of the platelet surface by quinine, (see Table 1). Anti-HuPlml (anti-Gpllla), FMC25 (anti-GptX) and AN51 (anti-Gplb) were labelled with 1251 and used in binding experiments with resting washed platelets in the presence a.nd absence of quinine. The binding of anti-HuPlml to platelets was not affected, FMC25 was only marginally inhibited by 1.O mM quinine, however AN51 binding was reduced by 35% and at 0.5 mM quinine 12% reduction in AN51 binding was observed. These studies were repeated on three occasions at several different dilutions with no significant inter assay variation.
Vol61, Nos. S/6
QUININE INDUCED THROMBOCYTOPENIA
509
Table 1 The Effect of Quinine and Drug Dependent Platelet Antibodies on the Binding of Radiolabelled Platelet Membrane Monoclonal Antibodies
Platelets Platelets plus 1 .O mM Quinine
Molecules of Mab boundlolatelet + (SD) AN51 FMC 25 HuPlml 4810 * 130 2960 f 110 3540 f 130 4860 + 100
2760 +
170*
2210 f
150*
Platelets plus 1 .O mM Quinine and 2860 & 100’ 2130 f 160 * 2630 f 120 ** DDAb Plasma 1 2720 f 80 1070 f 160 2860 f 90 ** DDAb Plasma 2 kDD 2900 f 50 PI ma3 A statistical evaluation using a paired t test was performed on the binding of Mabs to platelets (1) in the presence and absence of quinine, (2) treatment with quinine in the presence or absence of DDAB. [ The p c .05 * and p < .005 ** ] ??
??
??
inine nd DDab. Washed normal platelets were incubated with DDab plasma from three patients, or control plasmas in the presence of quinine prior to the addition of radiolabelled Mabs directed to platelet membrane glycoproteins. The binding of HuPlml to Gpllla was inhibited (30-40%) by prior binding of DDab in the presence of quinine and FMC25 binding, was also inhibited by 22% to 60%. However, AN51 binding was increased by approximately 30% compared with binding to platelets preincubated with quinine alone (Table 1).
Blockina
DISCUSSION The previous studies of the effect of quinine or quinidine on the platelet surface are incomplete and have relied primarily on the observation of functional effects. For example, Deykin and Hellerstein (18) demonstrated that quinidine inhibited platelet aggregation induced by ADP, collagen and adrenaline. The stereoisomers quinine and quinidine are generally considered to induce immune thrombocytopenia by similar mechanisms but it has been unclear whether quinine could affect the platelet surface directly causing perturbation of membrane glycoproteins and subsequent development of neoantigens or whether it behaves a hapten with the production of platelet antibodies. In the present study quinine both ex vivo and in vitro, is shown to inhibit platelet aggregation, induced by weak agonists e.g ADP and adrenaline, whereas aggregation induced by strong agonists e.g. collagen, thrombin, arachidonic acid and ristocetin was inhibited only in vitro in a dose dependent manner. Furthermore platelet aggregation induced by weak agonists ie adrenaline, ADP, low dose collagen, low dose thrombin were reversed by the addition of quinine after aggregation had commenced however aggregation induced by other agonists ie high dose collagen, arachidonic acid, ristocetin, thrombin 1U/ml was not affected. Although these effects were best demonstrated using quinine at greater than pharmacological concentrations there was an inhibitory effect on adrenaline induced aggregation within the range of concentrations used therapeutically. Platelets from normal
510
QUININE INDUCED THROMBOCYTOPENIA
Vol61, Nos. 516
volunteers taking quinine demonstrated an impaired response to stimulation with ADP, adrenaline and low dose collagen. This inhibitory effect could be overcome by increasing the concentration of the agonist in addition platelets appeared to respond normally to other agonists, e.g. thrombin, arachidonic acid. Platelet function returned to normal five days after cessation of quinine ingestion. Lawson et (19) have reported that quinidine administration (800 mg/day for 5 days) inhibited the platelet aggregation response to adrenaline. Quinidine ingestion also induced prolongation of the bleeding time from 4.0 f 0.5 minutes to 9.0 f 0.5 minutes without inducing thrombocytopenia. Thus demonstrating that in vitro quinine and quinidine effects on platelet function are not irrelevant phenomena and these effects may operate in vivo at therapeutic concentrations over several days. It is possible that a cumulative effect of quinine on platelet function at therapeutic concentrations for a prolonged period may have the effect equivalent to a higher concentration of drug for shorter exposure in vitro -Platelet aggregation induced by weak agonists is found to be impaired in patients with a release defect and it appears that released adhesive proteins are a necessary prerequisite for propagation of the platelet aggregate. Similarly Quinine not only inhibits fibrinogen binding but also impairs release thus it is probable that the inhibitory effect noted ex vivo on platelet aggregation induced by weak agonists involves a perturbation of the Gpllb-llla fibrinogen binding site or else precipitates a mild release defect. Data presented here also demonstrates that quinine inhibits exogenous fibrinogen and fibronectin binding to platelet receptors. This inhibition is not due to the competition from the release of endogenous proteins found within the a granule as the conditions employed in the ligand binding experiments did not elicit release of fibrinogen or fibronectin. These in vitro results were not generally demonstrated with volunteers’ platelets tested during quinine ingestion with the exception of impaired fibrinogen and fibronectin binding with one volunteer’s platelets ex vivo (results not shown). This impaired adhesive protein binding may be due to minor membrane changes induced by quinine occurring in vivo and in vitro. Mustard et (20) and Marguerie et al (21) have reported that on stimulation platelets undergo shape change, associated with the induction of specific fibrinogen receptors on the platelet surface and binding of fibrinogen resulting in the primary phase of aggregation. Quinine may, therefore, inhibit or alter the shape change and induction of specific fibrinogen receptors necessary for platelet aggregation. Quinine has no apparent effect once secondary irreversible aggregation has been initiated with strong agonists. We have also found that quinine caused the ‘elution’ of a substantial amount of fibrinogen previously bound to the platelet surface, but that bound fibronectin remains unaffected. Harfensit et al (22) have shown that radiolabelled fibrinogen bound to ADP stimulated platelets dissociated from those He also demonstrated that fibrinogen was platelets when they disaggregated. capable of dissociating from the platelet surface up to 10 minutes after it had bound to gel-filtered platelets stimulated by 10).1MADP. It is possible, therefore, that quinine induces some conformational alteration in the platelet membrane which is incompatible with fibrinogen binding thus preventing platelet aggregation. While the quinine-induced conformational changes also appear to prevent binding of fibronectin, once bound it may not dissociate possibly because of the presence of secondary binding sites which stabilise the linkage. The failure of quinine to dissaggregate platelets which have been stimulated with strong agonists may be due to the presence stable bridges of fibronectin cross-linking platelets in addition to fibrinogen bridging. (23) Other adhesive proteins have been shown to be important for platelet-platelet For example, Soria et (24), demonstrated that platelets from interaction. afibrinogenaemic patients responded to collagen or arachidonic acid in a normal
Vol61, Nos. 516
QUININE INDUCED THROMBOCYTOPENIA
511
fashion except that aggregates were smaller. Additionally a report by Dixit and coworkers (25) found that a monocional antibodies, raised against fibronectin, inhibited aggregation induced by strong agonists thus implicating fibronectin in the mechanism of platelet aggregation induced by these agonists. Several reports suggest that the glycoprotein complex Ilb-llla contains the platelet-fibrinogen receptor (26,27,28). Both quinine and anti-HuPlml block the binding of fibrinogen and fibronectin to Gpllla (29) but quinine does not inhibit the binding of anti-HuPlml to Gpllla indicating that quinine may effect adhesive protein binding at a site removed from the HuPlml epitope. Preincubation of platelets with quinine partially inhibited the binding of AN51 - a monoclonal antibody directed to Gplb - suggesting that quinine either blocked the epitopic site on Gplb or produced a conformational change in the platelet membrane so that the AN51 site was no longer accessible to the Mab either due to a decrease in the total number of available receptors or due to an effect on the association constant for binding of the Mab. Quinine had a slight effect on the binding of Mabs FMC25 to GplX and no effect of binding or HuPlml to Gpllla, indicating no alteration of this epitope. It appears that DDab and quinine together inhibit binding of HuPlml (antiGpllla) as well as FMC25 (anti-GplX) to the platelet surface whereas there is partial enhancement of AN51 (anti-Gplb a) binding compared with quinine treated platelets The inhibition demonstrated with HuPlml and FMC25 binding support the observations of Berndt et al (30) & Pfueller et al (3) who reported that the specificity of DDab was related to Gplb a and Gpllb-llla. However the interesting observation that AN51 binding is enhanced in the presence of DDab is consistent with a more accessible a chain of Gplb following quinine perturbation. We suggest that quinine causes widespread but specific conformational alterations in platelet membrane antigens which are associated with altered platelet function ex vivo and in vitro. These perturbations prevent the binding of fibrinogen and fibronectin therefore inhibiting platelet function, and may expose neoantigens against which a heterogenous population of quinine-induced antibodies may be directed.
ACKNOWLEDGMENTS Funds for the research were obtained from the National Health and Medical Research Council Grant. We thank Mrs. A. Scott for her secretarial assistance.
REFERENCES 1.
GRANDJEAN, L.C. A case of purpura haemorrhagica after administration of quinine with specific thrombocytopenia demonstrated in vitro Acta Med. &and. 727, (Suppl 213) 165-170, 1948
2.
LERNER, W., CARUSO, R., FAIG, D. & KARPATKIN, S. Drug-dependent and nondrug dependent antiplatelet antibody in drug-induced antiplatelet antibody in drug-induced immunologic thrombocytopenic purpura. Blood, 66, 306-311, 1985.
3.
PFUELLER, S.L., BILSTON, R.A., LOGAN, D., GIBSON,J.M. & FIRKIN, B.G. Heterogeneity of Drug-dependent platelet antigens and their antibodies in quinineand Quinidine-Induced Thrombocytopenia: Involvement of Glycoproteins lb, Ilb, Illa, IX. B/ood,72,1155-1162, 1988.
512
QUININE INDUCED THROMBOCYTOPENIA
Vol61, Nos. 5i6
4.
STRICHER, R.B. & SHUMAN, M.A. Quinidine purpura: Evidence Glycoprotein V is a target platelet antigen. Blood, 67, 1377-1381, 1986.
5.
KUNICKI, T.S., RUSSELL, N, NURDEN, A.T., ASTER, R.H. & CAEN, J.P. Further studies on the human platelet receptor for Quinine- and Quinidine- dependent antibodies. J. Immunol. 726, 398-402, 1981.
6.
CHRISTIE, D.J., MULLEN, P.C., & ASTER,R.H. Quinine- and quinidine platelet antibodies can react with Gpllb/lIla. Br. J. Haematol., 67, 213-219, 1987
7.
SHULMAN, N.R. lmmunoreactions 690,1958.
8.
SMITH, M.E., REID, D.M., JONES, O.E., JORDON, J.V., KAUTZ, C.A. & SHULMAN, N.R. Binding of quinine- and quinidine-dependent drug antibodies to platelets is mediated by the Fab domain of the immunoglobulin G is not Fc dependent. J. C/in. Invest., 79, 912-917, 1987.
9.
CHRISTIE, D.J., MULLEN, P.C., & ASTER, R.H. Fab-mediated binding of drugdependent antibodies to platelets in quinidineand quinine-induced thrombocytopenia. J. C/in./nvest., 75, 31 O-314, 1985.
10.
CHRISTIE, D.J., & ASTER, R.H. Drug-antibody-platelet interaction in quinineand quinidine-induced thrombocytopenia. J. C/in. Invest., 70, 989-998, 1982.
11.
CHRISTIE, D.J., WEBER, R.W., MULLEN, PC., COOK, J.M., ASTER, R.H. Structural features of the quinidine and quinine molecules necessary for binding of drug-induced antibodies to human platelets. J. Lab and C/in. Med.,704, 730740, 1984.
involving platelet.
that
J. Exp. Med., 707, 665
12. ONDER,O., WEINSTEIN, A. & HOYER, L. Psuedothrombocytopenia caused by platelet agglutinins that are reactive in blood anticoagulated with chelating agents.Blood, 56, 177-82, 1980. 13.
PEGELS, J.G., BRUYNERS, E.C.E., ENGELFRIET, C.P. & von dem BORNE, A.E.G.Kr. Pseudothrombocytopenia: An immunologic study on platelet antibodies dependent on ethylene diamine tetra-acetate. Blood, 59, 157-61, 1982.
14.
SMITH, I.L., & MARTIN, T.J. Platelet thromboxane synthesis and release reactions in myeloproliferative disorders. Haemostasis, 7 7, 119-l 27, 1982.
15.
HUNTER, W.M. & GREENWOOD, F.C. Preparation of t311-labelled hormone on high specific activity. Nature, 794, 495-496, 1962.
16.
CONNELLAN, J.M., THURLOW, P.J., BARLOW, B., LOWE, M. & MCKENZIE, I.F.C. Investigation of alternative mechanisms of collagen-induced platelet activation by using monoclonal antibodies to glycoprotein Ilb-llla and fibrinogen. Thrombo. Haemost. ,55, 153-l 57, 1986.
growth
Vol61, Nos. 516
QUININE INDUCED THROMBOCYTOPENIA
513
17.
CONNELLAN, J.M., QUINN, M. & WILEY, J.S. The use of 1251 labelled staphylococcal Protein A in the diagnosis of autoimmune thrombocytopenic purpura and other immune mediated thrombocytopenias. Pathology, 78, 11 l116,1986.
18.
DEYKIN, D. & HELLERSTEIN, L.J., The assessment of drug-dependent and isoimmune antiplatelet antibodies by the use of platelet aggregometry. J. Clh. Invest., 57, 3142-3153, 1972.
19.
LAWSON, D., MEHTA, J., MEHTA, P.,BFiADFORD C.L. & IMPERI G.A. Cumulative effects of quinidine and aspirin on bleeding time and platelet a2adrenoceptors: Potential mechanism of bleeding diathesis in patients receiving this combination. J. Lab. C/in. Med. 708, 581586, 1986.
20.
MUSTARD, J.F., PACKHAM, M.A., KINLOUGH-RATHBONE, R.L., PERRY, D.W. & REGOECZI, E. Fibrinogen and ADP-induced platelet aggregation. Blood, 52, 453-466, 1978.
21.
MARGUERIE, G.A., PLOW, E.G. & EDGINGTON, T.S. Human platelets possess an inducible and saturable receptor specific for fibrinogen. J. Rio/. Chem., 254, 5357-5363, 1979.
22.
HARFENIST, E.J., PACKHAM, M.A. & MUSTARD, J.F. Reversibilty of the Association of fibrinogen with rabbit platelets exposed to ADP.Blood, 56, 189198, 1980.
23.
LEUNG, L & NACHMAN, R. Molecular mechanisms Ann. Rev. Med., 37, 179-186, 1986.
of platelet aggregation.,
24. SORIA, J., SORIA, C., BORG, J.Y., MIRSHANI, M., PIGUET, H., TRON, P., Platelet aggregation occurs in congenital FESSARD, C. & CAEN, J.P. afibrinogenaemia despite the absence of fibrinogen or its fragments in plasma and platelets, as demonstrated by immunoenzymology. Br. J. Haematol., 60, 503514, 1985. 25. DIXIT, V.M., HAVERSTICK, D.M., O’ROURKE, K., HENNESSY, S.W., BROEKELMANN, T.J., MCDONALD, T.J., GRANT, J.A., SANTORO, S.A. & FRAZIER, W.A. Inhibition of platelet aggregation by a monoclonal antibody against human fibronectin. Proc. Nat/. Acad. Sci. USA, 82, 3844-3848, 1985. 26. BENNETT, J.S., VILAINE, G. & CINES, D.B. Identification of the fibrinogen receptor and human platelets by photoaffinity labelling. J. Rio/. Chem., 257, 8049-8054, 1982. 27. NACHMAN, R.L. & LEUNG, L.L.K. Complex formation of platelet membrane glycoproteins Ilb and llla with fibrinogen. J. C/in. Invest., 69, 263-269, 1982. 28. NACHMAN, R.L., LEUNG, L.L.K., KLOCZEWIAK, M. & HAWIGER, J. Complex formation of platelet membrane of glycoproteins Ilb and llla with the fibrinogen D domain. J. Biol. Chem., 259, 8584-8588, 1984.
514
QUININE INDUCED THROMBOCYTOPENIA
Vol61, Nos. 516
29.
THURLOW, P.J., KENNEALLY, D.A. & CONNELLAN, J.M. The role of fibronectin in platelet aggregation investigated with a monoclonal antibody Br. J. Haematol.,(ln press 1990).
30.
BERNDT, M.C., CHONG, B.H., BULL, H.A., ZOLA, H., CASTALDI, P.A. Molecular characterization of quinine/quinidine drug-dependent antibody platelet interaction using monoclonal antibodies. Blood, 66, 1292-l 301, 1985.
RESEARCH
61; 501-514,199l
0049-3848/91 $3.00 + .OOPrinted in the USA. Copyright
(c) 1991 Pergamon
Press pk. All rights reserved.
CHANGES IN PLATELET FUNCTION AND REACTIVITY INDUCED BY QUININE IN RELATION TO QUININE (DRUG) INDUCED IMMUNE THROMBOCYTOPENIA
J.M. Connellan, S. Deacon & P.J. Thurlow Department of Haematology, Austin Hospital, Studley Road, Heidelberg, Victoria, 3084, Australia (Received
7.9.1990;
accepted
in revised form 13.12.1990
by Editor H.H. Salem)
ABSTRACT Quinine, a drug known to induce immune mediated thrombocytopenia, has been postulated to mediate binding of drug dependent antibodies to a range of platelet membrane glycoproteins. Quinine may not act solely as a hapten however, as we have shown that it inhibits platelet aggregation (ex vivo and in vitro) and release and modifies the ability of activated platelets to bind the adhesive proteins fibrinogen and fibronectin in a dose dependent fashion. Studies on the effect of quinine on the binding of monoclonal antibodies HuPlml (Gpllla) FMC25 (GplX) and AN51 (Gplb) to platelets shows a selective reduction in AN51 binding. In addition quinine induced platelet antibodies from thrombocytopenic patients, in the presence of quinine, have been shown to inhibit binding of these monoclonal antibodies to platelets to varying degrees. These observations suggest that quinine causes widespread but specific conformational changes in platelet membrane antigens resulting in the production of quinine which may expose neoantigens induced antibodies.
INTRODUCTION Quinine and quinidine (cinchona alkaloid stereoisomers) have been demonstrated to induce anti-platelet antibodies resulting in severe thrombocytopenia in susceptible patients (1,2). Patients with quinine-induced thrombocytopenia have
Key words: Platelet antibody, quinine, thrombocytopenia, monoclonal antibodies to membrane glycoproteins. 501
fibrinogen,
fibronectin,
502
QUININE INDUCED THROMBOCYTOPENIA
Vol61, Nos. 516
been shown to possess a heterogeneous population of antibodies, with specificities for platelet glycoprotein (Gp) Ilb, Illa, lb, IX, V (3,4,5,6), which require the presence of the sensitizing drug for maximum binding of antibody to platelet. The initial studies (7) used to demonstrate the existence of quinine and quinidine antibodies relied upon platelet aggregometry where a platelet proaggregant effect was induced by adding patients’ serum together with quinine or quinidine to normal platelet rich plasma. However the effects of quinine or quinidine on the platelet membrane and the possible role this may play in the production of antibodies has been largely ignored. It has been assumed that quinine or quinidine act as haptens in the binding of antibodies to the platelet, by complexing with another moiety presumably either a plasma protein, or a protein on the platelet surface (8,9). However, no quinine-binding plasma factor has been demonstrated and Christie and co-workers (10,ll) have shown that quinine binds only weakly to platelet membranes. It is possible that quinine has an effect on the production of antibodies unrelated to its previously proposed role as a hapten. Quinine may have a direct effect in vitro and in vivo on platelet membrane proteins in which platelet structure and possibly function are effected. These changes could occur as a consequence of perturbation of membrane glycoproteins with concomitant exposure of cryptic antigens. In a similar fashion EDTA has been postulated to induce the exposure, in some patients, of a platelet cryptic antigen. Onder et al (12) demonstrated EDTA associated pseudothrombocytopenia is an h vitro phenomenon, and antiplatelet antibodies only react with platelets in the presence of EDTA (13). Similarly quinine mediated platelet antibodies may develop after exposure of a putative cryptic antigen in platelets caused by the presence of quinine. Our study was undertaken to investigate the effect of quinine (in vitro and ex W) on platelet structure and function as this may indicate the mechanism involvedin production of quinine induced antiplatelet antibodies. Our investigations demonstrate that quinine has marked effects on platelet function and also appears to directly alter membrane antigenicity thus supporting the concept that conformational alterations in platelet membrane antigens in the presence of quinine are important in the induction of quinine induced thrombocytopenia.
MATERIAL AND METHODS Quinine HCI was either modified Tyrode’s 2mM MgCL2) or Hepes KCI) or veronal buffered
purchased from ICN Pharmaceuticals Inc. and dissolved in buffer pH 7.4 (0.15 M NaCI, 2.6 mM KCI, 12 mM NaHC03, buffered saline pH 7.4 (10 mM Hepes, 0.15 M NaCI, 2.6 mM saline pH 7.4 as required.
Platelet rich plasma (PRP) was prepared from blood from either Platelet aaareaation. healthy donors who had not taken any medication in the previous two weeks or healthy volunteers who had ingested 600 mg of quinine bisulphate daily for four days. PRP, anticoagulated with sodium citrate (3.1%), was stored in a stoppered tube at 37” until used. Platelet aggregation studies were performed [l] prior to the ingestion of quinine, [2] whilst the subjects were ingesting quinine and [3] five days after cessation of quinine ingestion. The effect of quinine in vitro on platelets from healthy donors was studied by incubating 5Oyl of quinine (varying concentrations) or Hepes buffered saline with 400 ~1of PRP prior to the addition of 50 f_rlof agonist. Platelet aggregation was followed by using a Chronolog Lumi Aggregometer. Agonists used were (final concentrations): ADP 2, 4, 10 ,uM (Sigma) adrenaline 60 PM (Sigma), collagen 1.6, 8 pg/ml (Hormon-Chemie), arachidonic acid 500 yg/ml
Vol61, Nos. 516
QUININE INDUCED THROMBOCYTOPENIA
503
(Biodata), ristocetin 1 mg/mI (Wellcome) and thrombin 0.2 units/ml (Human plasma, Sigma). 146 serotonin (Amersham S.A. 57 mCi/m mole) was added Platelet release studies. to PRP (20 ~1-1 ).rCi/5 ml) and incubated for 30 minutes at 37”. PRP (400~1) was incubated with varying concentrations of quinine (50~1) prior to the initiation of aggregation with agonist (50~1); aggregation was monitored on a Chronolog Lumi Aggregometer; 250 ,ul of 4.09% EDTA was added to stop aggregation and the mixture centrifuged at 9000 x g for 5 minutes. The supernatant (500~1) was transferred to a scintillation vial together with 500 yl of distilled water and 10 ml of scintillation solution (toluene 2 parts, Triton x-100 1 part; PPO (2,5-diphenyloxazole) 4g/litre) (14) then mixed and counted. A blank (unstimulated, platelet pellet) and total (unstimulated platelets included) tubes were also counted. The percentage release of t4C-5HT was calculated as Test-Blank/Total Blank x 100. Human fibrinogen (Grade L) was obtained from Kabi (Stockholm) Adhesive proteins. and further purified by adsorption on gelatin-Sepharose GB-CL (Pharmacia) in order Fibronectin was purchased from Bethesda to remove fibronectin contamination. Research Laboratories (Maryland USA). Proteins were labelled with 125fodine using the chloramine T Protein labelling. method (modified from Hunter and Greenwood, 15) in brief; fibrinogen (1 mg), fibronectin or IgG (100 t.rg) were incubated with 51.11 of t25lodine 0.5 m Ci(Amersham) and 3 ~1 of chloramine T (lmg/ml); the mixture was incubated for between 2 and 7 minutes at room temperature, then 3yl of sodium metabisulphate (2.4 mg/mI) and 100 ~1 of potassium iodide (16mglml) were added and the mixture was incubated for between 5 and 15 minutes at room temperature. Labelled protein was separated from free iodine by chromatography on a PDlO column (Pharmacia). Gel filtered platelets were prepared from PRP after washing by Gel filtered platelets. centrifugation and resuspended in modified Tyrode’s buffer pH 6.4, 1% BSA. The platelets were separated from contaminating plasma by gel filtration chromatography on Sephadex 2B CL equilibrated with modified Tyrode’s buffer pH 7.4 1% BSA. Fibrinoaen bindina to activated olatelets. Gel filtered platelets (1.4 x 106) suspended in modified Tyrode’s buffer (with 5.5 mM glucose and 0.3% BSA but no MgC12) were preincubated with varying amounts of quinine (0 to 1 .O mM). Activated platelets were prepared by incubating collagen (1.6 yg/mI) with gel filtered platelets at 37” for 3 minutes. 125i fibrinogen was then added to the activated platelets and incubated for a further 30 minutes at 37” and platelet bound 1251 fibrinogen, in the platelet pellet, was measured after centrifugation through phthalate oil (S.G. 1.021). Fibronectin binding to activated olatelets . This was carried out using a similar method to that employed for fibrinogen binding with the exception that thrombin (1 unit/ml) was used to activate platelets and the modified Tyrode’s buffer contained 5.5 mM glucose and 1% BSA. Purification of IaG from plasmas of patients with drua dependent antibodies fDDab). Plasmas were applied to a Protein A Sepharose 66 CL column (Pharmacia), equilibrated with phosphate buffered saline pH 7.4, and incubated for 60 minutes at 4°C. Unbound proteins were eluted with phosphate buffered saline and 0.1 M glycine
504
QUININE INDUCED THROMBOCYTOPENIA
Vol61, Nos. 516
HCI pH 3.0 was used to elute bound IgG which was then dialysed against phosphate buffered saline and stored at -70°C. Purification of monoclonal antibodies (Mab) from ascites or serum. HuPlml, directed against Gpllla, was produced by the authors (16). AN51 directed to Gplb and FMC25 (ascites), directed to GplX, were gifts from Dr. M. Berndt, Westmead Hospital, N.S.W. Mabs were applied to Protein A Sepharose 6B CL column equilibrated with 0.1 M phosphate buffer pH 8.2, and incubated overnight at 4°C; unbound protein was eluted with phosphate buffer. Bound Mabs were eluted with 0.1 M citrate buffer pH 5.8 and dialysed against phosphate buffered saline pH 7.4. The purified Mabs were stored at -70°C at a concentration of 1 mg/ml. Drua Dependent Antibodies bindinq to platelets. Washed platelets were prepared from blood collected into EDTA and resuspended in Tyrode’s buffer pH 7.4. Platelets (1 x 108) were incubated for 30 minutes at 37°C with varying amounts of quinine (O1 .O mM) and DDab plasma or 300 mg of purified IgG (normal plasmas and IgG were used as controls). 1251 Staphylococcal Protein A (SPA) was added and incubated for 30 minutes at 37°C; platelet bound SPA was quantitated after the platelets were pelleted through phthalate oil (Density 1.021) (17). Effect of quinine on the bindinq of Mabs to platelets. Purified Mabs HuPIMI, AN51 and FMC25 were iodinated using the chloramine T method and used in binding experiments to platelets in the presence of varying concentrations of quinine using purified mouse IgG as a control. Washed platelets (1 x 108) resuspended in modified Tyrode’s buffer pH 7.4, 1% BSA were incubated with varying concentrations of quinine (O-l .O mM) and CaC12 (1 mM) following which dilutions of 1251 labelled Mabs were added and incubated for 30 minutes at 37°C. Platelet bound Mab was quantitated after the platelets were pelleted through phthalate oil.(See above) Blockinq of 1251 Mab bindina to platelets by preincubation of olatelets with DDab. Washed platelets (1 x 108) in Tyrode’s buffer pH 7.4, 1% BSA were incubated for 30 minutes at 37°C with plasma or purified IgG (300 r-19)(from patients with DDab or controls) in the presence of l.OmM quinine. Platelets were washed and incubated with 1251 labelled Mab for a further 30 minutes at 37°C. Platelet bound Mab was quantitated after the platelets were pelleted through phthalate oil.
RESULTS PRP, preincubated with Effect of auinine on platelet aaareaation and release. quinine at varying concentrations,was tested with a range of agonists in order to determine a dose response of quinine in platelet aggregation (See Fig 1). At 0.05 mM quinine, which is within the pharmacological range, minimal inhibition was observed only with adrenaline induced aggregation, this level of quinine did not affect aggregation induced by other agonists tested. (ADP (2,4 PM), collagen (1.6, 8.yg/ml), thrombin (0.2U/ml) and ristocetin (lmg/ml) results not shown). At 0.25 mM quinine, adrenaline induced aggregation was largely abolished and slight inhibition was observed in aggregation induced by other agonists. Higher concentrations of quinine (1 .OmM) effectively inhibited aggregation induced by all agonists tested and with 4 PM ADP slight reversible aggregation was demonstrated. Preincubation of PRP with varying amounts of quinine inhibited the release of 14CSerotonin from platelets activated with collagen. Serotonin release, induced by
Vol61, Nos. S/6
505
QUININE INDUCED THROMBOCYTOPENIA
low doses of collagen (1.6 pg/ml), was inhibited by 0.25 mM quinine, whereas 0.5 mM quinine was required to inhibit release induced by higher concentrations of collagen (8 pglml) (Fig. 2). C
B
A Adrenaline I~OJJM 1
Collagen 1Bug/ml)
QmM
ADP
QmM 0, 0.25
0
14!~M) QmM
005
o? O-25
0
12
3
4
0
5
12
3
4
5
(minutes)
Time
FIG. 1 Representative examples of the dose dependent effect of quinine on platelet aggregation induced by weak and strong agonists. PRP was incubated with varying amounts of quinine prior to the addition of agonists (A) adrenaline (60 PM), (B) collagen (8 pg/ml) or (C) ADP (4 PM) 70 I
0 0.0
I 0.2
I 0.4
I 0.6
I 0.8
Quinine (mM) FIG. 2 The effect of quinine on platelet release. 14CSerotonin loaded platelet were incubated with varying concentrations of quinine prior to the addition of collagen (0) = 1.6 pglml (0) = 8 pglml
506
Vol61, Nos. 516
QUININE INDUCED THROMBOCYTOPENIA
Aggregation of PRP initiated by most agonists was reversible when quinine (l.OmM) was added after aggregation had commenced. Aggregation induced by collagen 1.6yg, ADP 2,4 yM, thrombin 0.2 U/ml (data not shown) and adrenaline (60 PM) (Fig 38) were reversed by the addition of quinine. Whilst aggregation induced by collagen 8pg (Fig 3A) and ristocetin lmg/mI (not shown) were not affected by adding 1 mM quinine.
B
A Collagen
Adrenaline
(8Ng/ml)
(60 uM 1
Q (ImM)
/ 3
0
I
1
I
I
1
2
3
4
I
I
5
6
Time
I
I
0 (minutes
1
2
3
1
;
k
)
FIG. 3 Reversal of aggregation by quinine. Quinine (1 .O mM) was added during the course of (A) collagen (8 pg/ml) and (B) adrenaline (60 PM) induced platelet aggregation.
PRP collected from volunteers was studied in triplicate and the ingestion of quinine was found to diminish the aggregatory response to ADP, adrenaline and low dose collagen (1.6 pg/ml). Platelets exposed to quinine in vivo failed to respond to ADP 1pM (Fig 4A) and one out of three samples failed to aggregate when challenged with ADP 2yM whilst platelets from the other two volunteers showed a reduced response. Platelets from all volunteers showed diminished response (average 20% reduction) to ADP 4 I.IM which was a consistent finding. Similarly platelets tested during quinine ingestion, showed a diminished responsiveness to adrenaline, however, there was variation in the extent of aggregation in the individuals tested. Platelets from one volunteer did not respond to adrenaline 1.25 yM,(Fig 48) although aggregation did proceed in the other two volunteers it was accompanied by an extended lag period. All volunteers achieved therapeutic levels of quinine. Some reduction (average 30% decrease) in aggregation occurred after ingestion of quinine when low dose collagen (1.6 pg/mI) was used to stimulate platelets (Fig 4C) and this was overcome by increasing the concentration of collagen to 8 yg/ml (Fig 4D). The platelets responded normally to other agonists, i.e. thrombin, arachidonic acid and A23187. When platelets were tested 5 days after quinine
Vol6 1, Nos. 516
QUININE INDUCED THROMBOCYTOPENIA
507
ingestion was completed, the aggregation response to ADP, adrenaline and collagen returned to normal levels.
A
AOP
IlOpMi
ADRENALINE
B
i 1 25 uM
I
8!.;--”
40
Pre postQuinine
n
Qulnlne lnqestion
1
COLLAGEN
c
(1-61Jg/mli
I
I
0
0
1
2
3
4
5
6
I8 ug/mi I
COLLAGEN
801
Pre + Post 1 Quinine
,/, c
1
2 Trme
3
4
Im~nutes 1
0
1
2 Ttme
,
,
,
3
L
5
6
I mu-&es)
FIG. 4 Representative examples of the effect of quinine ingestion on platelet aggregation measured ex vivo prior to, during and 5 days after completion of ingestion of quinine (600 mg/day) Agonists employed were (A) 1 .O FM ADP, (B) 1.25 PM adrenaline, (C) 1.6 t.rg/ml collagen and (D) 8pg/ml collagen.
Effect of quinine on the bindina of adhesive oroteins to activated platelets. Gel filtered platelets when activated with either collagen or thrombin, without stirring, were induced to bind radiolabelled fibrinogen or fibronectin. Preincubation of platelets with varying amounts of quinine prior to the addition of agonist, resulted in the inhibition of binding of these adhesive proteins. Quinine (0.5 mM) almost completely abolished the binding of fibrinogen stimulated by collagen (1.6 pg/mI) (Fig 5), whereas higher concentrations of quinine (0.8 mM) appear to be necessary to block the binding of fibronectin to thrombin (lU/ml) stimulated platelets (Fig 6). When ImM quinine was added to platelets, which had previously been exposed to low dose collagen and labelled fibrinogen, 73% of the platelet bound fibrinogen was eluted. In contrast the level of bound labelled fibronectin was unaffected by quinine added 30 minutes after activated platelets had been exposed to fibronectin.
508
QUININE INDUCED THROMBOCYTOPENIA
0
i B .E .$
9 0.0
1
I
I
I
I
0.2
0.4
0.6
0.6
1.0
Quinine (mM)
0.0
0.2
0.4
0.6
Quinine (mM)
0.6
1.0
Vol61, Nos. 516
FtG. 5 The effect of quinine on the binding of fibrinogen to platelets. Gel filtered platelets were preincubated with varying amounts of quinine prior to the addition of agonists and labelled adhesive proteins. 1251 fibrinogen was added to collagen (1.6 ,ug/ml) stimulated platelets
FIG. 6 The effect of quinine on the binding of fibronectin to platelets. Gel filtered platelets were preincubated with varying amounts of quinine prior to the addition of agonists and labelled adhesive proteins. 1251 fibronectin was added to thrombin (1 unit/ml) stimulated platelets (0) or resting platelets (e)
The effect of quinine on the binding of Mabs directed to olatelet membrane glycoproteins. Mabs directed to platelet membrane glycoproteins were used to study possible perturbation of the platelet surface by quinine, (see Table 1). Anti-HuPlml (anti-Gpllla), FMC25 (anti-GptX) and AN51 (anti-Gplb) were labelled with 1251 and used in binding experiments with resting washed platelets in the presence a.nd absence of quinine. The binding of anti-HuPlml to platelets was not affected, FMC25 was only marginally inhibited by 1.O mM quinine, however AN51 binding was reduced by 35% and at 0.5 mM quinine 12% reduction in AN51 binding was observed. These studies were repeated on three occasions at several different dilutions with no significant inter assay variation.
Vol61, Nos. S/6
QUININE INDUCED THROMBOCYTOPENIA
509
Table 1 The Effect of Quinine and Drug Dependent Platelet Antibodies on the Binding of Radiolabelled Platelet Membrane Monoclonal Antibodies
Platelets Platelets plus 1 .O mM Quinine
Molecules of Mab boundlolatelet + (SD) AN51 FMC 25 HuPlml 4810 * 130 2960 f 110 3540 f 130 4860 + 100
2760 +
170*
2210 f
150*
Platelets plus 1 .O mM Quinine and 2860 & 100’ 2130 f 160 * 2630 f 120 ** DDAb Plasma 1 2720 f 80 1070 f 160 2860 f 90 ** DDAb Plasma 2 kDD 2900 f 50 PI ma3 A statistical evaluation using a paired t test was performed on the binding of Mabs to platelets (1) in the presence and absence of quinine, (2) treatment with quinine in the presence or absence of DDAB. [ The p c .05 * and p < .005 ** ] ??
??
??
inine nd DDab. Washed normal platelets were incubated with DDab plasma from three patients, or control plasmas in the presence of quinine prior to the addition of radiolabelled Mabs directed to platelet membrane glycoproteins. The binding of HuPlml to Gpllla was inhibited (30-40%) by prior binding of DDab in the presence of quinine and FMC25 binding, was also inhibited by 22% to 60%. However, AN51 binding was increased by approximately 30% compared with binding to platelets preincubated with quinine alone (Table 1).
Blockina
DISCUSSION The previous studies of the effect of quinine or quinidine on the platelet surface are incomplete and have relied primarily on the observation of functional effects. For example, Deykin and Hellerstein (18) demonstrated that quinidine inhibited platelet aggregation induced by ADP, collagen and adrenaline. The stereoisomers quinine and quinidine are generally considered to induce immune thrombocytopenia by similar mechanisms but it has been unclear whether quinine could affect the platelet surface directly causing perturbation of membrane glycoproteins and subsequent development of neoantigens or whether it behaves a hapten with the production of platelet antibodies. In the present study quinine both ex vivo and in vitro, is shown to inhibit platelet aggregation, induced by weak agonists e.g ADP and adrenaline, whereas aggregation induced by strong agonists e.g. collagen, thrombin, arachidonic acid and ristocetin was inhibited only in vitro in a dose dependent manner. Furthermore platelet aggregation induced by weak agonists ie adrenaline, ADP, low dose collagen, low dose thrombin were reversed by the addition of quinine after aggregation had commenced however aggregation induced by other agonists ie high dose collagen, arachidonic acid, ristocetin, thrombin 1U/ml was not affected. Although these effects were best demonstrated using quinine at greater than pharmacological concentrations there was an inhibitory effect on adrenaline induced aggregation within the range of concentrations used therapeutically. Platelets from normal
510
QUININE INDUCED THROMBOCYTOPENIA
Vol61, Nos. 516
volunteers taking quinine demonstrated an impaired response to stimulation with ADP, adrenaline and low dose collagen. This inhibitory effect could be overcome by increasing the concentration of the agonist in addition platelets appeared to respond normally to other agonists, e.g. thrombin, arachidonic acid. Platelet function returned to normal five days after cessation of quinine ingestion. Lawson et (19) have reported that quinidine administration (800 mg/day for 5 days) inhibited the platelet aggregation response to adrenaline. Quinidine ingestion also induced prolongation of the bleeding time from 4.0 f 0.5 minutes to 9.0 f 0.5 minutes without inducing thrombocytopenia. Thus demonstrating that in vitro quinine and quinidine effects on platelet function are not irrelevant phenomena and these effects may operate in vivo at therapeutic concentrations over several days. It is possible that a cumulative effect of quinine on platelet function at therapeutic concentrations for a prolonged period may have the effect equivalent to a higher concentration of drug for shorter exposure in vitro -Platelet aggregation induced by weak agonists is found to be impaired in patients with a release defect and it appears that released adhesive proteins are a necessary prerequisite for propagation of the platelet aggregate. Similarly Quinine not only inhibits fibrinogen binding but also impairs release thus it is probable that the inhibitory effect noted ex vivo on platelet aggregation induced by weak agonists involves a perturbation of the Gpllb-llla fibrinogen binding site or else precipitates a mild release defect. Data presented here also demonstrates that quinine inhibits exogenous fibrinogen and fibronectin binding to platelet receptors. This inhibition is not due to the competition from the release of endogenous proteins found within the a granule as the conditions employed in the ligand binding experiments did not elicit release of fibrinogen or fibronectin. These in vitro results were not generally demonstrated with volunteers’ platelets tested during quinine ingestion with the exception of impaired fibrinogen and fibronectin binding with one volunteer’s platelets ex vivo (results not shown). This impaired adhesive protein binding may be due to minor membrane changes induced by quinine occurring in vivo and in vitro. Mustard et (20) and Marguerie et al (21) have reported that on stimulation platelets undergo shape change, associated with the induction of specific fibrinogen receptors on the platelet surface and binding of fibrinogen resulting in the primary phase of aggregation. Quinine may, therefore, inhibit or alter the shape change and induction of specific fibrinogen receptors necessary for platelet aggregation. Quinine has no apparent effect once secondary irreversible aggregation has been initiated with strong agonists. We have also found that quinine caused the ‘elution’ of a substantial amount of fibrinogen previously bound to the platelet surface, but that bound fibronectin remains unaffected. Harfensit et al (22) have shown that radiolabelled fibrinogen bound to ADP stimulated platelets dissociated from those He also demonstrated that fibrinogen was platelets when they disaggregated. capable of dissociating from the platelet surface up to 10 minutes after it had bound to gel-filtered platelets stimulated by 10).1MADP. It is possible, therefore, that quinine induces some conformational alteration in the platelet membrane which is incompatible with fibrinogen binding thus preventing platelet aggregation. While the quinine-induced conformational changes also appear to prevent binding of fibronectin, once bound it may not dissociate possibly because of the presence of secondary binding sites which stabilise the linkage. The failure of quinine to dissaggregate platelets which have been stimulated with strong agonists may be due to the presence stable bridges of fibronectin cross-linking platelets in addition to fibrinogen bridging. (23) Other adhesive proteins have been shown to be important for platelet-platelet For example, Soria et (24), demonstrated that platelets from interaction. afibrinogenaemic patients responded to collagen or arachidonic acid in a normal
Vol61, Nos. 516
QUININE INDUCED THROMBOCYTOPENIA
511
fashion except that aggregates were smaller. Additionally a report by Dixit and coworkers (25) found that a monocional antibodies, raised against fibronectin, inhibited aggregation induced by strong agonists thus implicating fibronectin in the mechanism of platelet aggregation induced by these agonists. Several reports suggest that the glycoprotein complex Ilb-llla contains the platelet-fibrinogen receptor (26,27,28). Both quinine and anti-HuPlml block the binding of fibrinogen and fibronectin to Gpllla (29) but quinine does not inhibit the binding of anti-HuPlml to Gpllla indicating that quinine may effect adhesive protein binding at a site removed from the HuPlml epitope. Preincubation of platelets with quinine partially inhibited the binding of AN51 - a monoclonal antibody directed to Gplb - suggesting that quinine either blocked the epitopic site on Gplb or produced a conformational change in the platelet membrane so that the AN51 site was no longer accessible to the Mab either due to a decrease in the total number of available receptors or due to an effect on the association constant for binding of the Mab. Quinine had a slight effect on the binding of Mabs FMC25 to GplX and no effect of binding or HuPlml to Gpllla, indicating no alteration of this epitope. It appears that DDab and quinine together inhibit binding of HuPlml (antiGpllla) as well as FMC25 (anti-GplX) to the platelet surface whereas there is partial enhancement of AN51 (anti-Gplb a) binding compared with quinine treated platelets The inhibition demonstrated with HuPlml and FMC25 binding support the observations of Berndt et al (30) & Pfueller et al (3) who reported that the specificity of DDab was related to Gplb a and Gpllb-llla. However the interesting observation that AN51 binding is enhanced in the presence of DDab is consistent with a more accessible a chain of Gplb following quinine perturbation. We suggest that quinine causes widespread but specific conformational alterations in platelet membrane antigens which are associated with altered platelet function ex vivo and in vitro. These perturbations prevent the binding of fibrinogen and fibronectin therefore inhibiting platelet function, and may expose neoantigens against which a heterogenous population of quinine-induced antibodies may be directed.
ACKNOWLEDGMENTS Funds for the research were obtained from the National Health and Medical Research Council Grant. We thank Mrs. A. Scott for her secretarial assistance.
REFERENCES 1.
GRANDJEAN, L.C. A case of purpura haemorrhagica after administration of quinine with specific thrombocytopenia demonstrated in vitro Acta Med. &and. 727, (Suppl 213) 165-170, 1948
2.
LERNER, W., CARUSO, R., FAIG, D. & KARPATKIN, S. Drug-dependent and nondrug dependent antiplatelet antibody in drug-induced antiplatelet antibody in drug-induced immunologic thrombocytopenic purpura. Blood, 66, 306-311, 1985.
3.
PFUELLER, S.L., BILSTON, R.A., LOGAN, D., GIBSON,J.M. & FIRKIN, B.G. Heterogeneity of Drug-dependent platelet antigens and their antibodies in quinineand Quinidine-Induced Thrombocytopenia: Involvement of Glycoproteins lb, Ilb, Illa, IX. B/ood,72,1155-1162, 1988.
512
QUININE INDUCED THROMBOCYTOPENIA
Vol61, Nos. 5i6
4.
STRICHER, R.B. & SHUMAN, M.A. Quinidine purpura: Evidence Glycoprotein V is a target platelet antigen. Blood, 67, 1377-1381, 1986.
5.
KUNICKI, T.S., RUSSELL, N, NURDEN, A.T., ASTER, R.H. & CAEN, J.P. Further studies on the human platelet receptor for Quinine- and Quinidine- dependent antibodies. J. Immunol. 726, 398-402, 1981.
6.
CHRISTIE, D.J., MULLEN, P.C., & ASTER,R.H. Quinine- and quinidine platelet antibodies can react with Gpllb/lIla. Br. J. Haematol., 67, 213-219, 1987
7.
SHULMAN, N.R. lmmunoreactions 690,1958.
8.
SMITH, M.E., REID, D.M., JONES, O.E., JORDON, J.V., KAUTZ, C.A. & SHULMAN, N.R. Binding of quinine- and quinidine-dependent drug antibodies to platelets is mediated by the Fab domain of the immunoglobulin G is not Fc dependent. J. C/in. Invest., 79, 912-917, 1987.
9.
CHRISTIE, D.J., MULLEN, P.C., & ASTER, R.H. Fab-mediated binding of drugdependent antibodies to platelets in quinidineand quinine-induced thrombocytopenia. J. C/in./nvest., 75, 31 O-314, 1985.
10.
CHRISTIE, D.J., & ASTER, R.H. Drug-antibody-platelet interaction in quinineand quinidine-induced thrombocytopenia. J. C/in. Invest., 70, 989-998, 1982.
11.
CHRISTIE, D.J., WEBER, R.W., MULLEN, PC., COOK, J.M., ASTER, R.H. Structural features of the quinidine and quinine molecules necessary for binding of drug-induced antibodies to human platelets. J. Lab and C/in. Med.,704, 730740, 1984.
involving platelet.
that
J. Exp. Med., 707, 665
12. ONDER,O., WEINSTEIN, A. & HOYER, L. Psuedothrombocytopenia caused by platelet agglutinins that are reactive in blood anticoagulated with chelating agents.Blood, 56, 177-82, 1980. 13.
PEGELS, J.G., BRUYNERS, E.C.E., ENGELFRIET, C.P. & von dem BORNE, A.E.G.Kr. Pseudothrombocytopenia: An immunologic study on platelet antibodies dependent on ethylene diamine tetra-acetate. Blood, 59, 157-61, 1982.
14.
SMITH, I.L., & MARTIN, T.J. Platelet thromboxane synthesis and release reactions in myeloproliferative disorders. Haemostasis, 7 7, 119-l 27, 1982.
15.
HUNTER, W.M. & GREENWOOD, F.C. Preparation of t311-labelled hormone on high specific activity. Nature, 794, 495-496, 1962.
16.
CONNELLAN, J.M., THURLOW, P.J., BARLOW, B., LOWE, M. & MCKENZIE, I.F.C. Investigation of alternative mechanisms of collagen-induced platelet activation by using monoclonal antibodies to glycoprotein Ilb-llla and fibrinogen. Thrombo. Haemost. ,55, 153-l 57, 1986.
growth
Vol61, Nos. 516
QUININE INDUCED THROMBOCYTOPENIA
513
17.
CONNELLAN, J.M., QUINN, M. & WILEY, J.S. The use of 1251 labelled staphylococcal Protein A in the diagnosis of autoimmune thrombocytopenic purpura and other immune mediated thrombocytopenias. Pathology, 78, 11 l116,1986.
18.
DEYKIN, D. & HELLERSTEIN, L.J., The assessment of drug-dependent and isoimmune antiplatelet antibodies by the use of platelet aggregometry. J. Clh. Invest., 57, 3142-3153, 1972.
19.
LAWSON, D., MEHTA, J., MEHTA, P.,BFiADFORD C.L. & IMPERI G.A. Cumulative effects of quinidine and aspirin on bleeding time and platelet a2adrenoceptors: Potential mechanism of bleeding diathesis in patients receiving this combination. J. Lab. C/in. Med. 708, 581586, 1986.
20.
MUSTARD, J.F., PACKHAM, M.A., KINLOUGH-RATHBONE, R.L., PERRY, D.W. & REGOECZI, E. Fibrinogen and ADP-induced platelet aggregation. Blood, 52, 453-466, 1978.
21.
MARGUERIE, G.A., PLOW, E.G. & EDGINGTON, T.S. Human platelets possess an inducible and saturable receptor specific for fibrinogen. J. Rio/. Chem., 254, 5357-5363, 1979.
22.
HARFENIST, E.J., PACKHAM, M.A. & MUSTARD, J.F. Reversibilty of the Association of fibrinogen with rabbit platelets exposed to ADP.Blood, 56, 189198, 1980.
23.
LEUNG, L & NACHMAN, R. Molecular mechanisms Ann. Rev. Med., 37, 179-186, 1986.
of platelet aggregation.,
24. SORIA, J., SORIA, C., BORG, J.Y., MIRSHANI, M., PIGUET, H., TRON, P., Platelet aggregation occurs in congenital FESSARD, C. & CAEN, J.P. afibrinogenaemia despite the absence of fibrinogen or its fragments in plasma and platelets, as demonstrated by immunoenzymology. Br. J. Haematol., 60, 503514, 1985. 25. DIXIT, V.M., HAVERSTICK, D.M., O’ROURKE, K., HENNESSY, S.W., BROEKELMANN, T.J., MCDONALD, T.J., GRANT, J.A., SANTORO, S.A. & FRAZIER, W.A. Inhibition of platelet aggregation by a monoclonal antibody against human fibronectin. Proc. Nat/. Acad. Sci. USA, 82, 3844-3848, 1985. 26. BENNETT, J.S., VILAINE, G. & CINES, D.B. Identification of the fibrinogen receptor and human platelets by photoaffinity labelling. J. Rio/. Chem., 257, 8049-8054, 1982. 27. NACHMAN, R.L. & LEUNG, L.L.K. Complex formation of platelet membrane glycoproteins Ilb and llla with fibrinogen. J. C/in. Invest., 69, 263-269, 1982. 28. NACHMAN, R.L., LEUNG, L.L.K., KLOCZEWIAK, M. & HAWIGER, J. Complex formation of platelet membrane of glycoproteins Ilb and llla with the fibrinogen D domain. J. Biol. Chem., 259, 8584-8588, 1984.
514
QUININE INDUCED THROMBOCYTOPENIA
Vol61, Nos. 516
29.
THURLOW, P.J., KENNEALLY, D.A. & CONNELLAN, J.M. The role of fibronectin in platelet aggregation investigated with a monoclonal antibody Br. J. Haematol.,(ln press 1990).
30.
BERNDT, M.C., CHONG, B.H., BULL, H.A., ZOLA, H., CASTALDI, P.A. Molecular characterization of quinine/quinidine drug-dependent antibody platelet interaction using monoclonal antibodies. Blood, 66, 1292-l 301, 1985.
