Journal of' Clinical Immunology, Vol. 12, No. 6, 1992

Lupus Anticoagulant Antibodies Inhibit Collagen-Induced Adhesion and Aggregation of Human Platelets in Vitro I. OSTFELD, 1"4 N. DADOSH-GOFFER, 2 S. BOROKOWSKI, 1 J. TALMON, 1 A. MANI, 3 U. ZOR, 1 and J. LAHAV 2"5

Accepted: June 24, 1992

and point to an important role of external plasma membrane phospholipids, particularly PI, in collageninduced platelet activation.

The effect of circulating lupus anticoagulant on platelet interaction with collagen and other proteins was tested, with the aim of understanding the role of membrane phospholipids in platelet function. Plasma samples from 26 systemic lupus erythematosus (SLE) patients, containing circulating lupus anticoagulant (LAC), were examined for their effect on adhesion and aggregation of normal human platelets. We find that SLE plasma, but not normal plasma, inhibits platelet adhesion to collagen in a concentration-dependent manner. At a plasma concentration of 1% the inhibition was 73 -+ 9% (mean + SD). In sharp contrast, there was no effect on platelet adhesion to fibronectin. Purified IgG from the same plasma samples also had an inhibitory effect. At 15 p~g/ml (comparable in IgG concentration to 0.1% plasma) it inhibited adhesion to collagen by 33 -+ 11%. Inhibition could be abolished by preincubation of the LAC-containing plasma with cardiolipin (CL), phosphatidylinositol (PI), and, to a lesser extent, phosphatidylserine (PS) but not with phosphatidylchotine (PC) or phosphatidylethanolamine (PE). Inhibition could also be abolished by preincubation of the LAC-containing plasma with a 10-fold excess of washed normal platelets. The effect of 1% LAC plasma on platelet aggregation was as striking, showing 79 + 26% inhibition of collagen-induced aggregation, and it could also be abolished by preincubation of the LAC plasma with cardiolipin. In contrast, the effect of LAC plasma on thrombin-induced aggregation was rather modest. Our results indicate that antiphospholipid antibodies interfere with platelet adhesion and stimulation by collagen in vitro

KEY WORDS: Lupus anticoagulant; antiphosphotipid antibody; phosphatidylinositol; collagen receptor; ptatetet adhesion; platelet aggregation; thromboxane A2.

INTRODUCTION It is well-known that sera from systemic lupus erythematosus (SLE) patients having the lupustype circulating anticoagulant (LAC) (1, 2) contain autoantibodies that recognize phospholipids (PL) such as cardiolipin (CL), phosphatidylserine (PS), phosphatidylinositol (PI), and occasionally phosphatidylethanolamine (PE), but not phosphatidylcholine (PC) (3-7). It also contains antibodies recognizing DNA (8-13; for review see Refs. 10-12). These antibodies have been found in experimental animals too (11). PI and PS, located at the plasma membrane of the platelets (14-21), have an important role in several platelet functions induced by collagen and thrombin (2, 6, 14, 21). Thus the possibility was raised that LAC activity on platelet aggregation during blood coagulation is mediated via phospholipids (PL) located on the surface of the platelet. In another cell type we have previously demonstrated specific inhibition of exogenous phospholipase C activity on cells pretreated with the LAC antibody (22, 23). It was therefore of interest to study the effect of LAC on early events in platelet stimulatiom such as adhesion, thromboxane A 2 formation, and aggregation, and to investigate the specific role of the different PL in these phenomena. The present study shows that antibodies to PL are capable of iiahibiting (a) platelet adhesion to collagen but not to fibronectin and (b) aggregation

~Department of Hormone Research, The Weizmann Institute of Science, Rehovot 76100, Israel. 2Central Coagulation Laboratory, Beilinson Medical Center, Petah Tiqva 49100, Israel. 3Department of Hematology, Sheba Hospital, Ramat Gan, Israel. 4present address: Internal Medicine A, Beilinson Medical Center, Petah Tiqva, Israel. 5To whom correspondence should be addressed. 415

0271-9142/92/1109-0415506.50/0© 1992PlenumPublishingCorporation


induced by collagen but not thrombin. These activities are associated with the PL specificity of the antibody, being most effective with anti-PI activity.


Materials Soluble calf skin collagen (Worthington Biochemical Corp., Freehold, N J) was used in the adhesion assays. Collagen reagent (Horm Hormon-Chemie M/Jnchen GMBH, Prod. No. 10500) was used in the aggregation experiments. Thrombin was a kind gift from Dr. J. Fenton II, Division of Laboratories and Research, New York State Department of Health, Albany. Fibronectin was purified from human plasma over gelatin-Sepharose columns as described previously (24, 25). PL and cardiolipin were obtained from Sigma Chemical Company, St. Louis, MO. 3H-Thromboxane B 2 was purchased from New England Nuclear, Boston, MA. Methods LA C Plasma and IgG. Plasma was collected from 30 SLE patients, selected specifically on the basis of the presence of LAC activity. These included 21 females and 9 males, age (mean -+ SD) 40.6 -+ 14.4 years, followed in the Hematology Department of the Beilinson Medical Center and the Chaim Sheba Medical Center, in Israel. The LAC activity was examined according to Rosner et al. (1), using the recalcification time for screening and the kaolin clotting time mixture test (KCT) for determining the relative amount of the anticoagulant. The mean of the KCT index, as defined by Rosner et al. (1), was 33.4 +- 16.0% and the range was 15.2-62%. Twelve of the patients had SLE according to the criteria of the American Reumatology Association, eight had recurrent first- and second-trimester abortions, five had recurrent venous thrombophlebitis or arterial thrombosis, four had cerebrovascular accidents at a relatively young age, and one had pulmonary hypertension. The anti-DNA titer ranged from 5 to 90%, with a mean of 20 + 19%. The IgG component of the plasma was purified by affinity chromatography using a protein A-Sepharose column. The control normal human plasma (NHP) was obtained from six healthy volunteersm three males and three females--who did not take any medication during the month previous to dona-


tion of blood. Each control plasma was tested for KCT index and for its effect on platelet adhesion, prior to inclusion in the pool used as the control. The IgG component of the control pooled plasma was purified in the same way as the patients' IgG. Platelets. Fresh human blood was obtained by venipuncture from healthy volunteers who had not ingested any drug during the 10 days prior to bleeding. The blood was anticoagulated with 0.1 vol ACD anticoagulant (85 rrL~ trisodium citrate, 65 mM citric acid, 2% dextrose). The platelets were separated from proteins and other blood cells by differential centrifugation at 120g for 20 min at 22°C, followed by gel filtration as described before (24). Briefly, a column of Sepharose 2B was preequilibrated with Lindon buffer containing 129 mM NaCI, 8.9 mM NaHCO3, 2.8 mM KC1, 0.8 mM KH2PO4, 0.8 mM MgC12.6H20 , 2.4 mM CaClz, 10.9 mM Na citrate.2H20, 1 mg/ml glucose, I0 mM Tris, pH 7.3; the platelets were eluted with the same buffer. In preabsorption experiments where washed platelets were used, they were prepared by repeated (x3) centrifugation at 2300 rpm of PRP inhibited by 30 ng/ml PGE 1. The sediment was resuspended in the washing buffer (30 ng/ml PGEI, 2 mM EGTA, 144 mM NaCI, 4 mM KCI, 49 mM MgCI/'6H20, 12 mM NaHCO3, 0.4 mM NaH2PO4, 5.5 mM glucose). On the fourth time the sedimented platelets were suspended in Lindon's buffer. Platelet concentration was determined by counting in a hemocytometer. Phospholipids. PL stocks (PI, PS, PE, PC, CL) were dissolved and kept in a methanol/chloroform (95:5) solution. Just prior to use the solvent was evaporated under N 2 and Lindon buffer added, so that the final concentration of the PL was 1 ~g/ml. The solution was mixed thoroughly and sonicated for 5 min (Sanogen Automatic Cleaner, Bronson Europa, N.V., Netherlands), in order to form a suspension of PL micelles. Adhesion ofPlatelets. Platelet adhesion to collagen- and fibronectin-covered substrata was determined as described (24, 25). Briefly, round glass coverslips (Deckglase, 12 mm) were incubated for 4 h at 20°C in 24-well plates with 100 ~g/ml soluble calf skin collagen (Worthington), or 20 p~g/ml plasma fibronectin. After incubation, the coverslips were rinsed thoroughly in the Lindon buffer, and each was incubated with 300 I~l gel-filtered platelet suspension (5 x 107 cells/ml) for 45 rain at 37°C. Following another thorough rinse, the bound platelets were lightly fixed in 3.7% formaldehyde in phosphate-buffered saline (PBS) for 30 min and

Journal of Clinical Immunology, Vol. 12, No. 6, 1992



counted by direct visualization in a phase-contrast Zeiss photomicroscope III, using a reticule. The number of platelets adhering to collagen under these conditions is in the range of 400 platelets/mm 2 (24). In experiments in which the effect of LAC was to be determined, we used plasma samples of individual patients (not pooled) at concentrations of 0.11.0%, as specified for each type of experiment. This was compared to the effect of plasma pooled from six normal donors (NHP), after variations between donors was established (see Results). IgG was purified from the plasma of each patient and used without pooling at a concentration of 15 Ixg/ml. Normal human IgG, for comparison, was purified from the N H P and used at the same concentration as that of the patients. The LAC plasma, NHP, or IgG at the specified concentrations was preincubated for 45 min at 37°C with the platelets (5 x 107/ml), which were then added to the collagen- or fibronectin-coated coverslips. The effect of various free PL on LAC action was determined by preincubation of LAC plasma or protein A-purified IgG with the respective PL suspension for 45 min at 37°C prior to the final incubation with the platelets. Measurements o f TXA2 Production. Gel-filtered human platelets (2 x 108 cells/ml) were used for analysis of collagen-induced TXAz formation. The platelets in 0.5 ml of Lindon buffer were incubated with N H P or LAC (1%) for 1 hr at 37°C. Collagen, at various concentrations (0.1-3 txg/ml), was added for I0 min at 37°C. The incubation was terminated by centrifugation for 2 rain. The medium was collected and stored in -20°C for about a week before analysis by radioimmunoassay (RIA). Previous experiments showed that all the T X A 2 produced was released to the medium. TXA 2 RIA. Aliquots (10-100 txl) of the medium were incubated with monoclonal antibodies (Ab) directed against the stable product T X B 2. The dilution of the Ab was predetermined in order to achieve 40% binding. 3H-TXBz (7000 cpm/tube) was added and incubation was carried out overnight at 40C. Dextran (0.2 mg)-coated charcoal (2 mg) was added. The bound 3H-TXBz-Ab in the supernatant was separated by centrifugation and counted. Aggregation o f Platelets in the Presence of LAC. Collagen- and thrombin-induced aggregation of gelfiltered platelets suspended in Lindon buffer containing 1% bovine serum albumin (BSA) was measured using a four-channel Bio Data Aggregometer.

Journal of Clinical Immunology, Vol. 12, No. 6, 1992

The concentration of the collagen and thrombin was adjusted in each experiment to give 50-70% aggregation and ranged from 1 to 10 ~g/ml in the case of collagen and 0.05-0.1 U/ml in the case of thrombin. The effect of LAC plasma samples on platelet aggregation was determined by preincubation of the platelets with 1% LAC plasma samples or 15 lxg/ml LAC IgG samples for 30 min at 37°C. In parallel platelets were incubated with N H P (1%) or normal IgG (15 Ixg/ml) under the same conditions. RESULTS Plasma samples containing LAC were tested for their effect on the function of normal human platelets. Three aspects of platelet function were studied: adhesion to collagen or fibronectin substrata, T X A 2 production induced by collagen, and aggregation stimulated by collagen and thrombin. Adhesion First, the effect of normal plasma on platelet adhesion to collagen was established. In 13 experiments gel-filtered platelets obtained from healthy volunteers, when incubated with I% plasma prepared by pooling plasma from six healthy donors, showed adhesion that was 98 -+ 13% of the adhesion in the absence of plasma. The individual plasmas had a mean effect on platelet adhesion of 104 + 16% (SD; range, 8%134%). In all subsequent experiments the plasma pool was used as a reference and all data are in comparison to the effect of the plasma pool at the same concentration as that of LAC plasma. In order to establish the interexperimental variation between plasmas of the same individual, we tested four patients in five experiments each. We found that the mean interexperimental SD of each individual was 11.5% and never exceeded 15%. Thus repeated determinations of the same individual do not affect the mean one way or another. We next tested the effect of LAC-containing plasma on the adhesion of normal platelets to collagen. Eleven experiments were conducted using nine platelet donors. Three concentrations of LAC plasma were used: 0.1, 0.5, and 1.0%. We found that preincubation of platelets with individual LAC plasma inhibited platelet adhesion to collagencovered substrata in a concentration-dependent manner (Fig. 1), reaching inhibition as high as 73 -+ 9% (n = 8 patients) at a 1% final LAC plasma concentration. Repeated screening of all 26 pa-













Er~ 22 .E


LAC plosmo concentrofion {%) Fig, 1, Concentration dependence of the inhibitory effect of LAC plasma on platelet adhesion to collagen. Gel-filtered platelets were incubated with the indicated concentration of plasma at 37°C for 45 rain, followed by determination of platelet adhesion to collagen substrata, as discussed under Materials and Methods, Each data point represents the mean -+ SE obtained with plasma of 10 patients, a single determination per patient. A f3

"~ < 4C

iiiiiiii: III;IIII



o °-


Z: r-


lllI~l*ll *I~Iii!ii ::m:::l


fiHP LAC LAC NHP ~ C o l l a g e n ~ Fibronectin

Normal LAC IgG IgG ~Collagen~

Fig. 2. Inhibitory effect of (A) anti-PL antibodies (LAC) and (B) purified LAC IgG on adhesion of gel-filtered platelets to collagen and fibronectin substrate. Platelets were incubated for 45 min at 37°C with either 0.5% normal human plasma (NHP), 0.5% individual LAC plasmas, or 14 txg/ml purified IgG derived from pooled NHP or from individual LAC plasmas. Each column represents the mean -+ SE of determinations with individual (nonpooted) plasmas as described under Materials and Methods.

tients' plasma samples (2 or 3 determinations per patient, reaching a total of 63 determinations) at a 0.5% final plasma concentration showed a mean inhibition of 50 +-- 12% (range, 21-69%) relative to the adhesion in the presence of 0.5% pooled NHP or in the absence of plasma (Fig. 2A). In contrast, adhesion to surfaces coated with fibronectin, rather than collagen, was virtually not affected by LAC (Fig. 2A), When LAC plasma was preincubated with the collagen substratum for 45 min and the coverslip washed three times before adding platelets, there was no inhibition of platelet adhesion to the coverslips (data not shown), suggesting that the LAC antibody recognized an entity on the platelet surface rather than on the collagen molecule. IgG purified from the LAC plasma also inhibited platelet adhesion to collagen (Fig. 2B), confirming that inhibition of adhesion is due to the antibody present in the whole plasma. The question of the identity of the LAC antigen was further tested in a series of absorption experiments. Plasma samples from four patients were incubated with normal washed human platelets, prepared from freshly drawn blood, at final concentrations of 108 and 109 cells/ml. After 45 min of incubation at 37°C the plasma samples were centrifuged to remove the platelets and their effect on platelet adhesion was reevaluated at a concentration of 1% preabsorbed LAC plasma. We find that the inhibitory effect of the LAC plasma samples on platelet adhesion to collagen was completely abolished by absorption with high concentrations of normal platelets (Table I). Moreover, the KCT of the absorbed plasmas was also normal following this treatment (Table I), indicating that the immunoglobulins causing both effects bind to the surface of the normal platelet. Interestingly, the amounts of platelets required for absorption of the LAC plasma activities were 10-fold the normal platelet concen-

Table I. The Effect of Absorption of LAC Plasma with Increasing Amounts of Normal Human Platelets on Platelet Adhesion and on KCT Platelet-absorbed plasma Nonabsorbed plasma

10s cells/ml

109 cells/ml

Inhibition of adhesion

Inhibition of adhesion

KCT index

Inhibition of adhesion

KCT index

Patient no,


KCT index

1 2 3 4

58 59 52 52

55 31 26 17

55 69 53 55

33 37 0 1

ND -7 -2 -5

0 0 0 4






Journal of Clinical Immunology, VoL 12, No. 6, 1992




.~LAC 4c

~ 12-



~ 2c

® o

~NHP-CL: +


LAC . . . . . . . +

LAC ,.tgG -


Fig. 3. Prevention by cardiolipin (CL) of the inhibitory effect of LAC plasma or purified LAC IgG on platelet adhesion to collagen. One percent NHP, LAC, or IgG was preincubated with CL (1 p~g/ml) for 45 min at 37°C prior to final incubation with platelets. For details see the legend to Fig. 2.

tration in plasma. Absorption with normal platelet concentrations was hardly effective (Table I). We next tested the ability of exogenous PL to neutralize the effect of LAC antibody on platelet adhesion. LAC plasma of IgG was incubated with CL suspension for 30 min at 37°C and the mixture of LAC and CL was then added to the platelets for further incubation, followed by a platelet adhesion test. We found that CL abolished the effect of LAC and LAC IgG and restored platelet adhesion to normal levels (Fig. 3). Pooled NHP, whether or not preincubated with CL, had no effect on platelet adhesion. The specificity of the PL effect was tested by preincubating the LAC antibody with various acidic or basic PL. As shown in Fig. 4, the inhibitory effect of LAC plasma on platelet adhesion could be completely abolished by preincubation with PI and markedly reduced (75%) by PS. in contrast, PC or 6O-

< 40-

g 5~ 20






L AC ~ - - ~ " "

Fig. 4. Phospholipid specificity: reversal by particular PL of LAC inhibition of platelet adhesion to collagen. NHP (0.5%) or LAC (0.5%) was preincubated with PI, PS, PC, or PE (1 ixg/ml) as described in the legend to Fig. 3.

Journal of Clinical Immunology, Vol. 12, No. 6, 1992







Collagen ( H g / m l )

Fig. 5. Lack of effect of LAC antibodies on stimulation of TXA2 formation induced by collagen. Gel-filtered platetets were incubated with 1% pooled NHP or 1% individual LAC plasmas. Various concentrations of collagen (0.1-3 ~zg/ml)were added and TXA 2 formation was determined. The results are the mean +- SE of three experiments.

PE had no effect. A similar pattern of PL specificity was observed following preincubation of purified LAC IgG with the different PL (data not presented). TXA 2 Production

Collagen-induced stimulation of TXAz formation in the presence and absence of LAC is shown in Fig. 5. LAC had no inhibitory effect on TXAz formation, whether collagen was used in submaximal (0.1-0.3 p~g/ml) or nearly maximal (3 txg/ml)doses. LAC also was without any effect when the release of 3Harachidonic acid induced by similar concentrations of collagen was examined (data not shown). Aggregation

Collagen-induced platelet aggregation was markedly inhibited by LAC plasma (Fig. 6). i~eincubation of normal human platelets with I% LAC plasma caused marked inhibition of their collageninduced aggregation. In 108 determinations using plasma samples from 26 patients (each determined 4 times), preincubation of normal human platelets with 1% LAC plasma caused inhibition of 79 -_+26% (range, 32-92%), relative to aggregation in the presence of 1% NHP. NHP at a similar concentration did not affect aggregation. Thrombin-induced aggregation (28 determinations) was only modestly inhibited (25 + 24%) by LAC (Fig. 6). As in the case of adhesion, LAC effect on platelet aggregation was also mediated by PL. Preincubation of LAC with CL markedly suppressed the



100.2 r0 80131


Lupus anticoagulant antibodies inhibit collagen-induced adhesion and aggregation of human platelets in vitro.

The effect of circulating lupus anticoagulant on platelet interaction with collagen and other proteins was tested, with the aim of understanding the r...
1MB Sizes 0 Downloads 0 Views