Summary. Antibody-mediated platelet destruction is a poorly understood process, although several lines of evidence suggest that Fcγ receptor (FcγR)-expressing splenic macrophages may be involved. In this study, chemiluminescence (CL) was used to measure the in vitro metabolic response of human monocytes to platelets sensitized with a human immunoglobulin (Ig)G1 recombinant antihuman platelet antigen-1a (anti-HPA-1a) antibody (B2G1; P-hrIgG1). CL responses were inhibited, but not abrogated, in the presence of 10 µg/ml human IgG or murine IgG2a, suggesting that FcγRI was principally involved. Experiments to determine the effect of Fab fragments to FcγRII found that CL responses to P-hrIgG1 were significantly enhanced, indicating that crosslinking of monocyte FcγRII by platelet-bound hIgG may modulate concomitant activation by FcγRI. Several observations suggested that the CL responses to P-IgG were dependent on the activation of resting platelets during their co-culture with monocytes and their subsequent P-selectin-mediated adhesion. First, the magnitude of the CL response was related to the level of P-selectin expression following platelet activation with α-thrombin. Second, CL responses were inhibited in the presence of antibodies that block the binding of P-selectin to P-selectin glycoprotein ligand-1 but not when platelets were pretreated and then washed. Third, the addition of anti-HPA-1a to monocytes from HPA-1a-negative donors preincubated with HPA-1a-positive platelets resulted in rapid CL responses. Finally, PGI2 inhibited the CL response to resting P-hrIgG1. Thus, evidence is presented that the interaction of human monocytes with P-hrIgG1 is mediated by FcγRI, modulated via FcγRII, and enhanced by the presence of P-selectin on the platelet membrane.
Alloimmunization to platelet-specific antigens can cause feto-maternal alloimmune thrombocytopenia (FMAT), post-transfusion purpura (PTP) and refractoriness following platelet transfusion. In Caucasians, antibodies to the human platelet antigen-1a (HPA-1a) cause approximately 83% of cases of FMAT (Newman et al, 1989).
Antibody-mediated platelet destruction is a poorly understood process, although several lines of evidence suggest that Fcγ receptor (FcγR)-expressing splenic macrophages may be involved. In vitro, it has been demonstrated that monocytes are able to phagocytose and lyse platelets sensitized with anti-HPA-1a antibodies (Gengozian & Rice, 1982; Court et al, 1984). These interactions are competitively inhibited by fluid-phase immunoglobulin (Ig)G, suggesting that FcγR may mediate the recognition of sensitized platelets by mononuclear phagocytic cells (Barton et al, 1987; Saleh et al, 1989). In vivo, the administration of radiolabelled platelets to a patient with autoimmune thrombocytopenia (AIT) was reported to result in the accumulation of radiolabel in the spleen (McMillan et al, 1974). Moreover, splenectomy may be followed by a rise in the platelet count in some patients with AIT (McClure, 1975; Najean et al, 1997). Finally, the administration of an anti-FcγRIII antibody to a patient with autoimmune thrombocytopenia was followed by a rise in platelet count (Clarkson et al, 1986). Taken together, these results suggest that IgG-sensitized platelets may be destroyed in a manner analogous to the extravascular destruction of red cells.
The primary role of platelets is the maintenance of haemostasis by thrombus formation at sites of vascular damage. P-selectin is expressed at increased levels on activated platelets and, by binding to P-selectin glycoprotein ligand 1 (PSGL-1), P-selectin has been reported to mediate the slowing of leucocytes for firm adhesion at the sites of such damage (Jungi et al, 1986; Buttrum et al, 1993). Platelets have also been found to bind to monocytes in aP-selectin-dependent manner in whole blood (Rinder et al, 1991a,b; de Bruijne-Admiraal et al, 1992). Thus, it is possible that P-selectin might influence the IgG-mediated destruction of platelets by monocytes. To explore this possibility, we have developed a chemiluminescence (CL) assay to measure the metabolic response of human monocytes to platelets sensitized with anti-HPA-1a antibodies.
Materials and methods
Monoclonal antibodies. Unless stated otherwise, platelets were sensitized using a human (h) recombinant (r) IgG1 anti-HPA-1a (hrIgG1 anti-HPA-1a; clone B2G1, CAMTRAN007) prepared as described previously (Garner et al, 2000). For comparison, red cells were sensitized with hIgG1 anti-D (clone BRAD-5) obtained from the International Blood Group Reference Laboratory (IBGRL), Bristol, UK. hIgG1 anti-varicella zoster (clone VAZO-5; IBGRL) was used as a negative control.
Murine (m) IgG1 anti-CD41 (clone PM6/248) was obtained from Serotec, Oxford, UK. Murine IgG1 antiglycophorin A (extracellular domain; clone BRIC 256) was obtained from the IBGRL. Two mIgG1 anti-P-selectin antibodies were selected: Thromb-6 [Central Laboratory of the Netherlands Red Cross Blood Transfusion Service (CLB), Amsterdam, the Netherlands] blocks P-selectin-mediated binding of platelets to monocytes, and AK-6 (Serotec) binds P-selectin but does not inhibit this interaction (Skinner et al, 1991; De Bruijne-Admiraal et al, 1992). Serum-derived hIgG (Sigma, Poole, UK) and mIgG2a anti-glycophorin A (cytoplasmic domain; clone BRIC 163; IBGRL) were used to competitively inhibit FcγRI-mediated interactions, and Fab fragments of anti-FcγRII (clone IV.3; Medarex, Annadale, USA) were used to inhibit FcγRII-mediated responses.
Antisera. Three sera, donated with informed consent and containing anti-HPA-1a, were selected. PG1 was from a patient with post-transfusion purpura, and PG23 and PG165 were from mothers whose infants were affected by FMAT. Anti-human leucocyte antigen antibodies were not detected in these sera, which were heat inactivated (56°C, 20 min) before use. AB serum from an untransfused male donor was used as a negative control.
Red cell sensitization. Red cells (0·5 ml of a 3% v/v suspension of group O, R2R2 cells) were washed twice by repeated centrifugation at 700 g for 3 min and resuspension in 10 ml phosphate-buffered saline (PBS). Red cells were finally resuspended to 5% v/v (approximately 2 × 108/ml) in Roswell Park Memorial Institute (RPMI)-1640 medium (Gibco Invitrogen, Paisley, UK). Red cells (100 µl) were incubated with 100 µl of culture supernatant, containing 1 µg/ml hIgG1 anti-D or 5 µg/ml of mIgG1 anti-glycophorin A, for 1 h at 37°C with gentle agitation. Sensitized red cells (E-IgG) were then resuspended and washed three times with PBS as before. E-IgG were finally resuspended in Hank's balanced salt solution containing 0·5% bovine serum albumin (HBSS/BSA) to a final concentration of 1% v/v (approximately 4 × 107 cells/ml).
Platelet isolation and sensitization. Platelet-rich plasma (PRP) was prepared from whole blood collected with informed consent from normal donors into either di-potassium ethelendiaminetetraacetic acid (K2EDTA; Vacutainer™; Becton Dickinson, Luton, UK) or 10% v/v acid citrate dextrose (ACD; 2·5 g/100 ml tri-sodium citrate, 1·37 g/100 ml citric acid, 2 g/100 ml glucose). Blood was centrifuged for 10 min at 400 g, and PRP was collected as the upper fraction. Platelet concentration was measured using a Micro Diff II analyser (Coulter Electronics, Luton, UK).
In experiments to measure platelet binding IgG, PRP was diluted fivefold in PBS, containing EDTA (147 mmol/l NaCl, 26 mmol/l Na2HPO4, 12·6 mmol/l Na2EDTA) and 0·5% BSA (PBS/EDTA/BSA) (von dem Borne et al, 1978). Platelets were then collected by centrifugation (2000 g for 5 min), washed three times in PBS/EDTA/BSA and resuspended in 1 ml to a final concentration of 3 × 108/ml. Platelets were then incubated in the presence of 1 µg/ml hrIgG1 anti-HPA-1a or 5 µg/ml mIgG1 anti-CD41 or 5 µg/ml mIgG1 anti-glycophorin A (clone BRIC 256) for 30 min at 37°C. Sensitized platelets (P-IgG) were then washed three times with PBS/EDTA and resuspended in HBSS/BSA to a final concentration of 1 × 108/ml.
In functional assays, a method of platelet preparation and sensitization was required that minimized platelet activation. To achieve this, platelets were sensitized by adding hrIgG1 anti-HPA-1a (final concentration; 1 µg/ml) to PRP (3 × 108 platelets/ml) from ACD anticoagulated blood. Platelets were sensitized for 30 min at 37°C. One part PRP was then gently mixed with five parts modified Tyrodes salt solution (34 mmol/l NaCl, 12 mmol/l NaHCO3, 2·9 mmol/l KCl, 0·34 mmol/l Na2HPO4, 1 mmol/l MgCl2, 5 mmol/l glucose), containing 10 mmol/l HEPES, 10% ACD, 0·5% BSA, pH 7·4 (TSS/HEPES/ACD/BSA) in polypropylene tubes (Elkay, Basingstoke, UK). Platelets were pelleted by centrifugation (700 g, 10 min) and gently resuspended in TSS/HEPES/ACD/BSA. Platelets were washed a further two times before resuspension in TSS/HEPES/ACD/BSA or HBSS/BSA at 1 × 108/ml. The preparation was discarded if there was any visible sign of platelet aggregation or if more that 5% of platelets expressed elevated levels of P-selectin.
Determination of P-selectin expression and bound IgG. Expression of P-selectin was determined by flow cytometry. Platelets (10 µl at 1 × 108/ml) were incubated in round-bottomed, polypropylene tubes for 20 min in the dark with 10 µl of R-phycoerythrin (RPE)-conjugated mIgG1 anti-CD62P (clone AK-6) and 10 µl of fluorescein isothiocyanate (FITC)-conjugated goat antihIgG or antimIgG (Sigma). Samples were then resuspended in 500 µl of PBS/BSA and analysed by flow cytometry. In order to estimate the percentage of platelets expressing elevated levels of P-selectin, a gate was defined which excluded 95% of platelets in PRP from ACD anticoagulated blood.
Modulation of platelet activation. Sensitized platelets from ACD anticoagulated blood were activated using thrombin. Aliquots (66 µl) of washed P-IgG (1 × 108/ml) in PBS/BSA were dispensed into a U-well microplate and incubated with an equal volume of human α-thrombin (0–2·5 U/ml; Sigma) for 5 min. TSS/HEPES/ACD/BSA (150 µl) was added to each well and the microplate centrifuged for 5 min at 400 g. P-IgG were resuspended in TSS/HEPES/ACD/BSA at 1 × 108/ml. P-selectin expression was determined by flow cytometry, as described above. In all experiments, more than 90% of platelets expressed elevated levels of P-selectin following incubation with 2 U/ml α-thrombin.
Prostaglandin I2 was used to inhibit platelet activation. Washed P-IgG (1 × 108) were incubated with TSS/HEPES/ACD/BSA, containing 300 nmol/l prostaglandin I2 (Prostacyclin, PGI2; Sigma), for 10 min at room temperature. Platelets were then washed with TSS/HEPES/ACD/BSA and resuspended in HBSS/BSA at 1 × 108/ml.
In some experiments, washed P-IgG (1 × 108/ml) were incubated with 10 µg/ml of blocking or non-blocking mIgG1 anti-CD62P (Thromb-6 or AK-6 respectively) for 30 min at 37°C. Platelets were washed three times with TSS/HEPES/ACD/BSA and finally resuspended in HBSS/BSA at 1 × 108/ml.
The chemiluminescent response of monocytes to P-IgG and E-IgG. Chemiluminescence (CL) was used to measure the metabolic response of monocytes to P-IgG. Peripheral blood mononuclear cells (PBMC) were prepared by density gradient centrifugation (Histopaque 1077; Sigma) of EDTA anticoagulated blood from six random donors, unless stated otherwise. PBMC were then washed four times in PBS/BSA and resuspended at 3 × 105 monocytes/ml in HBSS containing 20% RPMI-1640 medium and 2% FCS (HBSS/RPMI/FCS). PBMC (100 µl) were added to white, opaque, flat-bottomed plates (Life Sciences International, Basingstoke, UK) and left to adhere for 2 h at 37°C, in a humidified atmosphere of 5% CO2. Wells were then washed once with HBSS/BSA to remove unbound mononuclear cells and 100 µl of HBSS/BSA was added to each well. Then 50 µl of HBSS, containing luminol (40 mmol/l), and 50 µl of P-IgG or E-IgG were added to each well. Chemiluminescence was recorded using an Anthos Lucy 1 (Labtech, Ringmer, UK). Maximal monocyte CL responses generally occurred 15 or 30 min after the addition of P-IgG or E-IgG respectively. CL responses were, therefore, calculated as the sum of 1 s readings taken every 2·35 min over these periods. In every experiment, the background CL generated in the presence of unsensitized platelets or red cells was subtracted from the CL generated in the presence of sensitized cells.
In some experiments, monocyte FcγRs were blocked by incubating adherent monocytes for 20 min at 37°C with 50 µl of monomeric hIgG (0–400 µg/ml), monomeric mIgG2a (0–100 µg/ml) or Fab fragments of anti-FcγRII (0–20 µg/ml) prior to the addition of P-IgG or E-IgG.
Typing of the H131/R131 polymorphism on FcγRII. Allele-specific primers to the polymorphism at amino acid residue 131 of FcγRIIA were used in hybridization assays, as described by Osborne et al (1994).
The biological activity of hrIgG1 anti-HPA-1a
A recombinant anti-HPA-1a antibody was used throughout these studies to eliminate the possibility of contaminating antibodies, to reduce experimental variability and to avoid platelet exposure to serum proteins. To determine if the hranti-HPA-1a was functionally similar to anti-HPA-1a antibodies from human serum, the monocyte CL response to platelets sensitized with hranti-HPA-1a was compared with the response to platelets incubated with different antisera. Figure 1 demonstrates that there was no significant difference in the CL responses obtained. Furthermore, there was a direct relationship between the magnitude of the CL response and the amount of platelet-bound antibody.
Effect of blockade of monocyte FcγR on the CL response to P-IgG
Crosslinking of monocyte FcγRI by red cells sensitized with IgG anti-D leads to the generation of superoxide by monocytes; the CL responses which result are abrogated in the presence of approximately 10 µg/ml monomeric hIgG or mIgG2a (Hadley et al, 1988). The ability of fluid-phase hIgG to abrogate the CL response to E-hIgG1 was confirmed in the current study (Fig 2). In contrast, the CL response to P-hrIgG1 was not abrogated by up to 80 µg/ml monomeric IgG. At 40 µg/ml of monomeric IgG, the inhibition of the CL response to P-IgG1 (88 ± 6%) was significantly lower than the inhibition of the CL response to E-IgG1 (99 ± 1%, P < 0·0017). Concentrations of monomeric IgG greater than 100 µg/ml resulted in non-specific CL responses. Similar results were obtained using mIgG2a (data not shown).
The inability to abrogate CL responses to P-hrIgG1 by blocking FcγRI raised the possibility that FcγRII might be involved (Klaassen et al, 1990). To test this possibility, monocytes were incubated with Fab fragments of a monoclonal antibody (clone IV.3) with specificity for FcγRII. The treatment of monocytes from donors homozygous for the polymorphism R131/R131 with Fab fragments of IV.3 inhibited the FcγRII-mediated CL response to platelets sensitized with mIgG1 anti-CD41 (Fig 3). In marked contrast, the CL response to P-hrIgG1 was enhanced by prior treatment with Fab fragments of IV.3. In a similar way, Fab fragments of IV.3 inhibited the CL response to red cells sensitized with mIgG1 antiglycophorin A but enhanced the response to red cells sensitized with hIgG1 anti-D (Fig 3).
Taken together, these data suggest that the inability of monomeric IgG to abrogate the monocyte CL response to P-IgG was not due to the presence of an FcγRII-mediated response.
Effect of P-selectin expression on the CL response to P-IgG1
Platelets express membrane proteins, which promote platelet–monocyte interactions. P-selectin has been shown to mediate the initial phases of platelet–leucocyte interactions. Experiments were, therefore, conducted to examine the role of P-selectin in the generation of a CL response to hrP-IgG1.
First, the effect of platelet isolation and sensitization on P-selectin expression was determined by measuring the binding of RPE-conjugated anti-CD62. Platelets isolated from EDTA anticoagulated blood or subjected to relatively high centrifugation speeds expressed higher levels of P-selectin than those isolated from ACD anticoagulated blood [mean channel fluorescence = 1·02 ± 0·5 (P = 0·032) vs 1·58 ± 0·3 (P = 0·012) vs 0·68 ± 0·5 respectively]. Platelets from ACD anticoagulated blood were used in subsequent experiments. Sensitization of platelets with hrIgG1 anti-HPA-1a did not further increase P-selectin expression (mean channel fluorescence = 0·57 ± 0·1).
To determine the effect of P-selectin expression on monocyte CL responses, P-hrIgG1 was incubated with up to 2·5 U/ml α-thrombin. Unexpectedly, P-hrIgG1 exposed to low concentrations of α-thrombin (0·01 U/ml) and expressing ‘intermediate’ levels of P-selectin (CD62int) elicited markedly reduced CL responses (Fig 4) compared with those elicited by untreated P-hrIgG expressing low levels of P-selectin (CD62lo) or P-IgG treated with high concentrations of α-thrombin and expressing high levels of P-selectin (CD62hi) (Fig 4). Similar results were obtained in several experiments summarized in Table I. The ‘restoration’ of CL responses following P-hrIgG treatment with relatively high concentrations of α-thrombin was not due to an effect on the level of antibody binding (Table I) nor to an effect of α-thrombin on the monocytes because the CL responses to E-IgG1 were unaffected (data not shown).
Table I. The relationship between P-selectin expression following incubation of P-hrIgG with α-thrombin and the magnitude of the CL response.
Thrombin concentration (U/ml)
CD62P expression (MFI)
anti-HPA-1a binding (MFI)
CD62P expression (MFI)
anti-HPA-1a binding (MFI)
CD62P expression (MFI)
anti-HPA-1a binding (MFI)
P-hrIgG1 or unsensitized platelets were incubated with 0–2·5 U/ml α-thrombin and then washed and added to monocytes. P-selectin expression and CL responses were then measured as described above. Results show the range of CL responses elicited by P-hrIgG, arbitrarily designated as CD62lo, CD62int and CD62hi according to mean fluorescence intensity (MFI). Three donors were used in three separate experiments.
The role of P-selectin expression in eliciting CL responses to P-IgG was tested using two mIgG1 anti-CD62 antibodies: Thromb-6 which blocks P-selectin mediated binding and AK-6 which does not. When P-IgG1 expressing high levels of P-selectin were incubated with 10 µg/ml of Thromb-6 and then washed, CL responses were inhibited by approximately 66%(Table II). In contrast, AK-6 did not reduce the CL responses (Table II). This suggested that the P-selectin that was expressed on activated P-IgG enhanced the monocyte CL response. Interestingly, the pretreatment of ‘resting’ P-hrIgG with Thromb-6 did not reduce the CL response. In contrast, when ‘resting’ P-hrIgG were incubated with monocytes in the presence of Thromb-6, CL responses were reduced by approximately 70%. This suggested that ‘resting’ platelets might be induced to degranulate and express P-selectin during their incubation with monocytes in the CL assay. Consistent with this hypothesis was the observation that treatment of ‘resting’ P-hrIgG1 with 300 nmol/l (PGI2) (before their addition to the monocytes) reduced the CL response by 45% ± 13% (range 29% to 58%, P < 0·0001, four experiments performed in quadruplicate). In comparison, the CL response to E-IgG1 was unaffected by 300 nmol/l PGI2 (data not shown).
Table II. Effect of anti-CD62P on the CL response to P-hrIgG1.
CL responses to thrombin-activated platelets pretreated with anti-CD62P or PBS
CL responses to untreated platelets pretreated with anti-CD62P or PBS
CL responses to untreated latelets in presence of anti-CD62P
P-hrIgG or unsensitized platelets were treated with 0·5 U/ml of α-thrombin or PBS and then incubated with 15 µg/ml of blocking (clone Thromb-6) or non-blocking (clone AK-6) anti-CD62P or PBS for 20 min. Washed or unwashed platelets were then added to monocytes and CL responses were measured as described above. Data shown are the mean ±1 SD.
Donor 1 (n = 3)
44·5 ± 5·6
40·7 ± 7·5
10·5 ± 3·0
39·4 ± 4·3
40·9 ± 3·8
43·4 ± 5·2
41·8 ± 6·9
13·8 ± 3·1
Donor 2 (n = 2)
83·6 ± 10·1
80·2 ± 9·3
32·1 ± 3·5
79·8 ± 7·5
79·2 ± 6·2
82·0 ± 5·0
80·5 ± 8·5
31·9 ± 3·8
Donor 3 (n = 3)
120·0 ± 10·3
110·5 ± 10·9
48·3 ± 5·3
100·6 ± 12·3
106·2 ± 11·4
103·9 ± 10·3
98·9 ± 10·8
47·6 ± 6·2
The pre-adhesion of platelets to monocytes affects the CL responses to P-IgG
P-selectin has been reported to promote the initial tethering of platelets to leucocytes (Jungi et al, 1986). The notion that the adhesion of platelets to monocytes might similarly enhance the CL response to P-hrIgG was, therefore, explored. The CL response of monocytes from HPA-1a-negative donors to P-hrIgG1 was compared with the response obtained by preincubating monocytes with unsensitized platelets followed by the addition of hranti-HPA-1a. Figure 5 represents a typical experiment showing that the peak CL response to P-hrIgG1 was achieved after approximately 12 min. When h-IgG1 anti-HPA-1a was added to monocytes immediately following the addition of HPA-1a-positive platelets, the peak CL was observed after approximately 15 min. However, when platelets were first allowed to adhere for 10 min, the peak CL was obtained approximately 4 min after the addition of h-IgG1 anti-HPA-a. Taken together, these results support the idea that P-selectin might mediate the adherence of P-hrIgG1 to monocytes in a manner that increases the antibody-mediated CL response.
The current study has used CL to measure the interaction between P-IgG and human monocytes in an assay intended to reflect the immune destruction of platelets by splenic macrophages. Results are presented which suggest that this antibody-dependent interaction is mediated by FcγRI, modulated via FcγRII, and enhanced by the presence of P-selectin on the platelet membrane.
The role of FcγRI and FcγRII in mediating the monocyte CL response to P-IgG
The ability of monomeric hIgG or mIgG2a to inhibit the CL response to P-hrIgG suggested that FcγRI was principally involved in mediating the antibody-dependent interaction with human monocytes. In a similar way, hIgG and mIgG2a have been shown to competitively inhibit the FcγRI-mediated CL response of monocytes to IgG anti-D sensitized red cells (Hadley et al, 1988).
Experiments to determine the effect of Fab fragments to FcγRII on these responses yielded unexpected results; CL responses to both E-hIgG1 and P-hrIgG1 were significantly enhanced, although the FcγRII-mediated responses toE-mIgG1 and P-mIgG1 were inhibited as predicted. This suggested that crosslinking of monocyte FcγRII by red cell or platelet-bound hIgG may downregulate concomitant activation by FcγRI. This downregulation might be mediated by FcγRIIb, which has recently been identified on human monocytes (Pricop et al, 2001). This receptor contains an inhibitory cytoplasmic signalling motif (ITIM) which downregulates B lymphocytes (Van den Herik-Oudijk et al, 1995; Karlsson et al, 1999).
The role of P-selectin in mediating the interaction of P-IgG with human monocytes
Blocking experiments showed that while the monocyte CL response to E-hIgG1 was abrogated in the presence of approximately 10 µg/ml hIgG or mIgG2a, it was not possible to totally inhibit the monocyte response to P-hrIgG1 under similar conditions. The inability to increase the level of inhibition by the addition of Fab fragments to FcγRII (discussed above) suggested that P-hrIgG binding through FcγRII was probably not responsible for this residual activity. We, therefore, speculated that P-selectin might mediate the adherence of platelets to monocytes in an FcγR-independent manner, resulting in the exclusion of monomeric IgG from regions where the platelet and monocyte membranes are brought into close proximity. This would allow high-affinity interactions between FcγRI and platelet-bound IgG even in the presence of competing fluid-phase IgG. Several observations were consistent with this. First, thrombin-activated platelets expressing high levels of P-selectin (CD62hi) elicited greater CL responses than platelets expressing lower levels of P-selectin (CD62int). Second, CL responses were inhibited by the presence of antibodies that block the binding of P-selectin by PSGL-1. Third, the addition of anti-HPA-1a to monocytes from HPA-1a-negative donors preincubated with HPA-1a-positive platelets resulted in rapid CL responses.
Two observations suggested that the adhesion of resting P-hrIgG (CD62lo) to monocytes might follow their activation during co-culture. First, CL responses were inhibited in the presence of Thromb-6 but not when platelets were pretreated and then washed. Second, PGI2 inhibited the CL response to resting P-hrIgG1 (but not E-IgG). It is not clear why P-hrIgG, treated with low concentrations of α-thrombin, did not activate in the assay and elicit CL responses comparable to those elicited by resting platelets or those activated by higher concentrations of α-thrombin. However, platelets treated with low concentrations of α-thrombin have been reported to possess reduced functional activity, paralleled by a reduced expression of P-selectin (Lozano et al, 1997). Similar observations have been made for other platelet agonists (Simard-Duquesne, 1973; Holme et al, 1977). This partial activation has been reported to result in reduced platelet aggregation under flow conditions (Aursnes et al, 1987; Savion et al, 2001).
Monocytes incubated with unsensitized thombin-activated platelets did not generate increased levels of CL over approximately 15 min. However, increased CL responses were observed after approximately 30 min (data not shown). This in agreement with Nagata et al (1993) who showed that monocytes generate superoxide after prolonged culture with thrombin-activated platelets.
The P-selectin-mediated adherence of P-IgG to monocytes or macrophages may be clinically significant in vivo. Rinder et al (1991a,b) showed that approximately 85% of monocytes from peripheral blood had one or more platelets bound. Moreover, leucocytes are intimately involved in thrombus formation and clearance (Celi et al, 1994; Kirchhofer et al, 1997). The current study presents data that suggest that haemostatic events might contribute to the destruction of IgG-sensitized platelets. It follows that events that result in platelet activation, such as sepsis, hypertension and storage for transfusion, might exacerbate the immune destruction of platelets (Rinder et al, 1991c; Andrioli et al, 1996; Konijnenberg et al, 1997).
We are grateful to Dr Belinda Kumpel (Bristol Institute for Transfusion Sciences, Bristol, UK) for isolating DNA and to Dr J. G. J van der Winkel (Department of Immunology, University Hospital Utrecht, the Netherlands) for performing FcγRII allotyping. We thank Dr Kathryn Armour (Department of Pathology) and Dr Willem Ouwehand (Division of Transfusion Science, University of Cambridge, and NBS) for the clone B2G1. We also thank Dr Rosey Mushens (IBGRL, Bristol) for antibody production and purification.