Development of a high-throughput ELISA assay for platelet function ...

392

© Schattauer 2010

New Technologies, Diagnostic Tools and Drugs

Development of a high-throughput ELISA assay for platelet function testing using platelet-rich plasma or whole blood Isabelle I. Salles1,2; Katleen Broos1; Alexandre Fontayne1; Tímea Szántó1; Changgeng Ruan3; Alan T. Nurden4; Karen Vanhoorelbeke1; Hans Deckmyn1 1Laboratory

for Thrombosis Research, KU Leuven-Campus Kortrijk, Kortrijk, Belgium; 2Department of Haematology, Imperial College London, London, UK; 3 Jiangsu Institute of Hematology, The First Affiliated Hospital of Soochow University, Suzhou, China; 4Centre de Référence des Pathologies Plaquettaires, Plateforme Technologique et d'Innovation Biomédicale, Hôpital Xavier Arnozan, Pessac, France

Summary Platelets play an essential role in the development of cardiovascular diseases and are the target of several agents that can inhibit their function. Despite the existence of a wide array of techniques to study platelet function, an assay to evaluate several platelet signalling pathways in a high-throughput fashion, combined with minimal blood volume and handling is still needed. We have developed a sensitive assay in the form of a sandwich ELISA where monoclonal antibodies against P-selectin or αIIbβ3 and GPIbα were used to capture and detect platelets, respectively, in the presence of five different agonists [ADP, TRAP (thrombin receptor agonist), U46619 (thromboxane A2 analogue), collagen-related-peptide, and arachidonic acid]. Binding of platelets to the antibodies increased dose-dependently with the concentration of

Correspondence to: Isabelle Salles Department of Haematology, Imperial College London Hammersmith Campus, Commonwealth Buildung DuCane Road, London, W12ONN, UK Tel.: +44 2083832298, Fax: +44 2083832296 E-mail: [email protected]

Introduction Activation of platelets plays a fundamental role in the physiology of haemostasis, and the pathophysiology of thrombosis and bleeding. Once vascular integrity is compromised, and subendothelial components (e.g. collagen, von Willebrand factor [VWF], fibronectin, laminin) become exposed, platelets undergo a highly regulated set of processes including adhesion, spreading, and aggregation in order to restore vascular integrity. In the high shear arterial system, platelets slow down by interacting reversibly to collagen-bound VWF via the platelet receptor glycoprotein (GP)Ib/V/IX, a process also known as “rolling” (1). Once the platelets have decelerated on VWF-collagen, they firmly adhere to collagen via the main collagen receptors GPVI and integrin α2β1, triggering a critical signalling that results in firm platelet adhesion to the injured vessel wall, and in the formation of a platelet monolayer (2, 3). Platelet activation occurs through various signal transduction pathways that elevate intracellular calcium. The GPVI receptor acts through tyrosine kinases triggering an activation cascade that Thrombosis and Haemostasis 104.2/2010

either agonist, while binding of ADP-activated platelets was abrogated when inhibitors of platelet activation were concomitantly added. The test showed good sample reproducibility in 15 healthy donors with conserved platelet response to agonists throughout the assay. Healthy subjects could be identified as normal-, hypo- or hyper-responders for each agonist, which for most cases (73%) was confirmed upon retesting. Finally, we demonstrated that the platelet ELISA assay can not only be used in platelet-rich plasma but also in whole blood; it now awaits large scale studies to assess its full screening and diagnostic values.

Keywords Platelet function tests, platelet ELISA, platelet activation

Received: July 31, 2009 Accepted after major revision: March 19, 2010 Prepublished online: May 27, 2010 doi:10.1160/TH09-07-0505 Thromb Haemost 2010; 104: 392–401

recruits various signalling molecules, including phospholipase Cγ (PLCγ) and phosphoinositol 3-kinase (PI3-kinase) (4). Activation of PLC leads to an increase of Ca2+ and to the activation of protein kinase C (PKC) responsible for platelet granular secretion releasing soluble agonists such as ADP, ATP, and serotonin that together with synthesised thromboxane A2 (TXA2), result in an amplification loop of platelet activation through receptors that activate PLCβ via G-proteins. (5). Concomitant activation of the coagulation cascade leads to local generation of the strong agonist thrombin, which on the platelets cleaves protease-activated receptors (PAR) 1 and 4 (6). The endpoint of the signalling cascade is activation of integrin αIIbβ3 (GPIIb/IIIa) that now can bind fibrinogen resulting in cross-linking of the platelets and the final platelet aggregate. Additionally, αIIbβ3 can bind VWF, further contributing to integrin-mediated thrombus formation (7). Finally, platelet activation is limited to the site of vessel damage by agents that increase either cAMP (such as prostacylin, prostaglandin D2 and E1, adenosine) or cGMP (such as nitric oxide [NO]) (8, 9). A variety of techniques to measure platelet function ex vivo have been developed (10–18). Overall, the gold standard for detecting

Salles et al. Platelet-activation ELISA test

platelet function by a given agonist remains light transmission aggregometry in platelet-rich plasma (PRP), although this technique is time-consuming, requires a relative large volume of blood, varies in sensitivity and is not high-throughput. Whole blood aggregometry on the other hand has the advantage of requiring smaller quantities of blood and less handling, but faces the same other drawbacks. Recently, a number of point-of-care tests such as PFA100, VerifyNow® or PlateletWorks® were introduced, which are fast and easy to perform, but give a relative broad view of overall platelet and ligand functionality, and offer limited refined dissection of the specific platelet pathways involved. Finally, an impressive study of platelet responses on 506 volunteers was carried out by flow cytometry, but focused on responses to adenosine diphosphate (ADP) and collagen-related peptide (CRP) platelet agonists only (10). We now report an assay that allows simultaneous screening of different activation and inhibitory platelet signalling pathways, paving the way for high-throughput screenings in healthy individuals and in patients.

Materials and methods Platelet agonists / antagonists ADP, TRAP, MRS-2179, arachidonic acid (AA) and prostaglandin E1 (PGE1) were from Sigma (St-Louis, MO, USA). TXA2 analogue U46619 and NO-donor S-nitroso-N-acetyl penicillamine (SNAP) were obtained from Calbiochem (San Diego, CA, USA). Collagen related peptide (CRP) was a generous gift from Dr. R. Farndale, (University of Cambridge, Cambridge, UK). Aspirin (ASA; L-Lysine Acetylsalicylic Acid) was from Sanofi-Aventis (Brussels, Belgium).

Monoclonal antibodies (MoAbs) MoAbs anti-P selectin (CD62P) SZ51, anti-GPIbα 6B4 and antiαIIbβ3 16N7C2, used in this study were produced and characterised as previously described (19–21). CD62P-PE was from Becton Dickinson Bioscience (San Jose, CA, USA). Fab fragments of moAb 6B4 were biotinylated using NHS LC Biotin (sulfosuccinimidyl 6-(biotinamido)hexanoate), according to the manufacturer’s instructions (Pierce, Rockford, IL, USA).

Platelet preparation Blood was drawn from healthy volunteers who had not taken any medication for the past two weeks, into sodium citrate (0.011 M) by free flow. PRP was prepared by centrifugation at 200 g for 10 minutes (min) at room temperature (RT). Platelet-poor plasma (PPP) was prepared by centrifugation at 1500 g for 6 min. Platelet © Schattauer 2010

count was determined with Cell-Dyn 1300 (Abbott-laboratories, Abbott Park, IL, USA) and adjusted to 200 x 103 platelets/μl with autologous PPP. For the study of platelet responses, blood was collected from 15 healthy donors during a week-long testing period (phase 1). Subjects were recalled three weeks later for a second phase of testing (phase 2) using the same methods as in phase 1.

ELISA Wells of 96-well microtiter plates (Greiner, Frickenhausen, Germany) were coated overnight at 4°C with 100 μl/well of moAb 16N7C2, or SZ51 (5 μg/ml in phosphate-buffered saline [PBS]). Plates were blocked with 3% milk powder in PBS (250 μl/well) for 2 hours (h) at RT. ADP, TRAP, U46619, CRP and AA were prepared by a serial dilution in PBS with starting concentrations of 20 μM, 40 μM, 30 μM, 125 ng/ml, and 400 μM, respectively, in a final volume of 25 μl and were added to the moAb-coated wells. In order to study the inhibition of platelet function, antagonists of platelet activation, SNAP, PGE1 and MRS-2179 were also prepared by serial dilutions in PBS (starting concentration 2 mM, 5 μM, and 100 μM, respectively) in the presence of a constant concentration of ADP (10 μM) in a final volume of 25 μl. PRP incubated with or without freshly made ASA (80–100 μg/ ml) for 10 minutes at RT, or whole blood (75 μl each) was added to the wells containing the different platelet agonists or antagonists (25 μl) and incubated at RT for 30 min. To detect captured platelets, biotinylated moAb 6B4 was used at 1 μg/ml in 0,3% milkPBS and was incubated for 1 h at RT, followed by a 1 h incubation with horseradish peroxidase-coupled streptavidin (Roche, Mannheim, Germany) (1/5,000 0.3% milk-PBS, 1 h at RT). After each incubation, wells were washed three times with PBS, except when whole blood was used (number of washes was increased to 5 after initial removal of blood). The colour reactions were initiated by adding H2O2 and orthophenylenediamine (OPD; Sigma) and stopped with 4 M H2SO4. Platelet binding was evaluated by measuring the absorbance at 490 nm on a microplate reader (EL340, Bio-Tek instruments, Winooski, VT, USA). Background values from uncoated wells, or coated wells incubated without platelets were subtracted from each reading. For each agonist/antagonist concentration, duplicate measurements were carried out.

Flow cytometry Platelets (1.106) pre-treated with or without ASA (0.1 or 1 mg/ml) were activated with 20 μM ADP or 400 μM AA in the presence or absence of various inhibitors (PGE1, SNAP, MRS-2179) for 10 min at RT. Platelets were then incubated with moAb SZ51 (0.2 μg/ml) and goat anti-mouse-PE (0.1 mg/ml; Jackson Immunoresearch, Suffolk, UK) each for 15 min at RT. Samples were then diluted in PBS and analysed on a FACScan flow cytometer (EPICS® XLMCL, Beckman Coulter, Fullerton, CA, USA). A gate analysis on Thrombosis and Haemostasis 104.2/2010

393

394


10,000 events was performed to select platelets based on their forward and side scatter properties.

Statistical analysis Agreement between duplicate measurements at different agonist concentrations was assessed by Lin’s concordance correlation that is used to assess the concordance in two continuous data sets (10, 22). Values of ±1 denote perfect concordance and discordance while a value of zero denotes its complete absence. All other correlation coefficients were quantified by the Spearman rank correlation test using GraphPad Prism 5.0 (GraphPad Software INC, San Diego, CA, USA). Correlation coefficients of 0 < r < 1 or –1 < r < 0 indicate that the two variables tend to increase or decrease together or that one variable increases as the other decreases, respectively. Correlation coefficients were regarded as significant if p < 0.05. Curve fitting to platelet activation/inhibition, calculations of EC50/IC50, coefficients of variation (CV) and confidence intervals

(CI) were done using GraphPad Prism 5.0. When EC50 could not be determined, set values corresponding to the highest or lowest agonist concentration were given according to the ratio of activation (corresponding to OD490[max agonist] /OD490[min agonist]) 1.3, respectively, in order to proceed for statistical evaluation. Lower and upper 99 % CI of the mean were calculated for the 15 healthy subjects during phase 1 and 2, allowing the discrimination of high and low responders, respectively.

Results Platelet activation by various agonists measured in the ELISA assay In this new test to assess platelet function, we especially focused on the high throughput aspect allowing for different stimuli as well as several donors to be tested simultaneously. A sandwich ELISA assay in a 96-well plate format was developed where resting and activated states of platelets could be distinguished. To this end, we se-

A

B

C

D

Figure 1: Binding of platelets to anti-CD62P and anti-αIIbβ3 moAb is induced by different agonists. MoAbs SZ51(●) or 16N7C2 (❍) (5 μg/ml) were coated in a 96-well microtiter plate overnight at 4°C and subsequently incubated with platelets (PRP), isolated from healthy subjects, in the presence of 0–20 μM ADP (A), 0–30 μM U46619 (B), 0–40 μM TRAP (C) and Thrombosis and Haemostasis 104.2/2010

0–125 ng/ml CRP (D). Platelet binding was detected by addition of biotinylated-6B4 moAb (anti-GPIb)/streptavidin-HRP. Sigmoidal regression fit was applied to the curves and EC50 calculated using GraphPad software. All data points represent duplicate measurements (mean ± standard error [SE]) from a representative donor (n ≥ 10).

© Schattauer 2010


lected moAbs to capture platelets that recognise platelet receptors present in relatively high numbers e.g. αIIbβ3 or P-selectin and more importantly that are up-regulated upon platelet activation, albeit to a lesser extent for the former (23, 24). The anti-GPIbα moAb 6B4 (25) was chosen to reveal the binding of platelets as it does not differentiate between resting and activated platelets, and furthermore as GPIb is present in fairly high numbers on the platelet surface (25,000 copies) allowing for a robust signal (26). Activation with various agonists of platelets isolated from healthy donors led to an increase in platelet binding to the anti-P selectin or anti-αIIbβ3 moAbs (씰Fig. 1). The specificity of the test was confirmed by using platelets from a Glanzmann’s thrombasthenia (GT) patient (27). As expected, no platelet binding could be detected in the ELISA assay using the 16N7C2 moAb under any of the conditions (data not shown). For healthy donors, plotting the OD (490 nm) versus the concentration of ADP, U46619, TRAP or CRP resulted in binding curves which could be fitted by sigmoidal regression (Fig. 1A-D). Doses of agonists for which a half maximal P-selectin expression are obtained (EC50) could subsequently be calculated for ADP (2.7 ± 0.4 μM), U46619 (7.6 ± 0.3 μM),

Inhibition of platelet function in the ELISA assay Next, we investigated whether inhibitory pathways could also be studied, for which we used SNAP and PGE1 that by inducing cGMP and cAMP generation, respectively, inhibit platelet activation and aggregation. Binding of platelets to moAb SZ51 after ADP activation (10 μM) could be prevented by addition of SNAP in a dose-dependent manner as illustrated in 씰Figure 2A, with an IC50 value of 0.88 ± 0.2 mM. Similarly, PGE1 limited activationdependent expression of P-selectin by ADP in platelets of a repre-

A

B

C

D

Figure 2: Binding of platelets to anti-CD62P moAb is inhibited by SNAP and PGE1. Platelet responses from two healthy donors were monitored in parallel in the platelet ELISA assay (A, C) or by flow cytometry (B, D). A, C) SZ51-coated microtiter plates were incubated with PRP in the presence of variable concentrations of antagonist SNAP 0–2 mM (A) or PGE1 0–5 μM (C), and a fixed concentration of ADP (10 μM). Platelet binding was detected by addition of biotinylated moAb 6B4 and streptavidin-HRP. Sigmoidal re© Schattauer 2010

TRAP (10.2 ± 0.2 μM) and CRP (7.9 ± 0.2 ng/ml). Similarly, when the anti-αIIbβ3 moAb was used in the ELISA test, EC50-values could also be calculated for ADP (0.9 ± 0.5 μM), U46619 (1.1 ± 0.5 μM), TRAP (19.9 ± 0.2 μM) and CRP (7.6 ± 0.3 ng/ml) (Fig. 1A-D). Responses to weaker agonists (ADP and U46619) were not as marked as for the anti-P selectin moAb, and therefore the anti-P selectin moAb SZ51 was chosen for subsequent experiments.

gression fit was applied to the curves and IC50 calculated using GraphPad software. Each data point represents mean of duplicate measurements ± SE from a representative donor (n ≥ 3). B, D) Platelet P-selectin exposure after activation of 10 μM ADP in the presence of 2 mM SNAP (B) or 2 μM PGE1 (D) was assessed by FACS as described in Materials and methods. Values represent the percentage of CD62P-positive platelets (representative donor, n ≥ 3).

Thrombosis and Haemostasis 104.2/2010

395

396


sentative healthy donor with an IC50 value of 2.6 ± 0.2 μM (Fig. 2C). These results were consistent with data obtained with both inhibitors using flow cytometry (Fig. 2B, D). Several flow cytometry studies have demonstrated that ASA did not affect platelet activation assessed by P-selectin expression when agonists other than AA were used (28–30). We therefore next evaluated the sensitivity of the platelet ELISA test to ASA. As expected, when the ELISA assay was performed with PRP (± ASA) in combination with ADP, U46619, TRAP or CRP and using the SZ51 moAb as a coating agent, no significant difference in EC50 or ratios of activation could be detected between the two conditions (data not shown). Although no significant differences in EC50 between PRP treated with ASA compared to control PRP were seen when AA was used as an agonist, a decrease in the ratios of activation could be observed (씰Fig. 3A-B). This was corroborated by similar results obtained in flow cytometry using ADP, U46619, TRAP, CRP and AA in combination with an anti-human CD62P Ab where only an effect of ASA was observed for AA (Fig. 3C; and data not shown).

Finally, we assessed the sensitivity of the ELISA to an inhibitor of the P2Y1 receptor. As illustrated in Figure 3D for a representative donor, addition of MRS-2179 in the ELISA prevented the full activation of platelets with 10 μM ADP (IC50 = 25.2 ± 0.3 μM). Similar effects of the MRS-2179 inhibitor on ADP-mediated platelet activation were also detected using flow cytometry as previously reported (31) (Fig. 3E).

Assessment of inter- and intra-variability The reproducibility of the test was evaluated by analysing platelet responses to the various agonists from 15 healthy individuals that were tested twice three weeks apart (phase 1 and 2). Reproducibility of technical replicates was measured using Lin’s concordance correlation (22) and was > 0.86 for each agonist for both experiments (씰Table 1). Individual’s response for the different agonists

C

B

A

D

Figure 3: Platelet ELISA is sensitive to ASA and MRS-2179. Platelet responses from two healthy donors were monitored in parallel in the platelet ELISA assay (A, B, D) or by flow cytometry (C, E). A, B, D) SZ51-coated microtiter plates were incubated with PRP from a healthy donor incubated without (●) or with 100 μg/ml ASA (❍) with indicated concentrations of AA (0–400 μM) (A). PRP from a healthy donor was incubated with MRS-2179 (0–100 μM) in the presence of 10 μM ADP (D). Platelet binding was detected by addition of biotinylated moAb 6B4 and streptavidin-HRP. Sigmoidal regression fit was applied to the curves and EC50 or IC50 calculated using GraphPad Thrombosis and Haemostasis 104.2/2010

E software. Each data point represents the mean of duplicate measurements ± SE from a representative donor (n ≥ 3). B) Ratios of activation were calculated as described in Materials and methods (n ≥ 3). C, E) Platelet P-selectin exposure after activation of 400 μM AA in the presence or not of ASA (C) or after activation of 20 μM ADP in the presence of 100 μM MRS-2179 (E) was assessed as described in Materials and methods. Values represent the percentage of CD62P-positive platelets (representative donor, n ≥ 3).

© Schattauer 2010


was assessed in each phase by calculating the Spearman’s correlation coefficients between adjacent doses for each agonist and between ratios of activation (씰Tables 2 and 3). Correlations between doses and between ratios of activation were high (r > 0.61; p < 0.05), attesting that individual’s platelet response was conserved throughout the assay in each phase (Tables 2 and 3). Furthermore, for each agonist, platelet activation, assessed as the mean of the ratios of activation or EC50, was not significantly different in phase 1 and 2 (씰Fig. 4A and B, p > 0.05). Overall responses to each agonist were reproducible over time. However, when analysing individual data, responses varied considerably between the normal individuals as illustrated by the high CV observed for each agonist and each phase (씰Table 4) and EC50-values from phase 1 not correlating with EC50-values from phase 2 (data not shown). We next determined the lower and upper 99% CI of the mean for each agonist, to classify healthy donors as low-, normal- and high-responders (Fig. 4B-E) for phase 1 and 2. Categories showed good reproducibility as the same category was assigned in 73% of the cases, of which a majority of individuals tested normal in the two visits. Interestingly, for both phases, two individuals were identified as hypo-responders to ADP or U46619 (data not shown) while two others were hyper-responsive to ADP (data not shown) or TRAP (Fig. 4E).

Use of the platelet ELISA assay with whole blood As our test does not rely on light transmission, platelet activation could also be analysed in whole blood which obviously requires less handling and volume, and may be preferable for platelets as they are in a more physiological environment. Platelet responses to different agonists were therefore assessed in whole blood in parallel with PRP. The sigmoidal regression curves calculated with PRP and whole blood data concurred with each other as illustrated for a representative donor (씰Fig. 5A-D; n = 4). Expectedly, EC50 values obtained in whole blood for each agonist correlated very well with those in PRP with a Spearman’s correlation coefficient r = 0.81 (p < 0.001; n = 4).

Discussion Determination of platelet function classically is done by using platelet aggregometry based on changes in light transmission in PRP, originally described by Born (32), which has proven its usefulness during the years, as it closely mimics the physiological process where several stimuli induce a whole activation cascade that finally results in αIIbβ3-mediated platelet aggregation. However, this technique suffers from several drawbacks: although the results are obtained “in real time”, it takes around 5 min to have a full aggregation, which in view of the relative short active ex vivo life time of platelets, precludes study of larger numbers of (ant)agonists at different concentrations, unless a battery of aggregometers is used in © Schattauer 2010

parallel. The reproducibility of the test is not optimal, the interpretation is not always straightforward (which parameter: lag time, slope, maximal amplitude…), and depends on the machine used: 250–1,000 μl of PRP is needed, although technology has improved in this area by the development of computerised multichannel aggregometers, also available for whole blood aggregation. In addition, in contrast to the obvious detection of hypo-active platelets (e.g. GT, Bernard-Soulier syndrome [BSS] and other congenital disorders, or due to the use of anti-platelet drugs) it has so far been particularly hard to determine hyper-active platelets, although an elaborate FACS-analysis clearly provided evidence that within a healthy population, subjects with low and high responding platelets can be detected (10). Other parameters, such as secretion of ATP (luminescence) or appearance of platelet activation markers [activated integrins, expression of granular proteins such as P-selectin (CD62P), lysosome integral membrane protein-3

Table 1: Analysis of the variability of the data in the assay: low variability between duplicates. Lin’s concordance correlation between two replicates for each agonist is indicated for each experiment with seven duplicates/agonist/donor (n = 15). Values close to 1 indicate high concordance. Agonist

Phase 1

Phase 2

ADP

0.88

0.86

U46619

0.92

0.93

TRAP

0.86

0.93

CRP

0.91

0.93

Table 2: Analysis of the variability of the data: good concordance between adjacent dosages. Spearman’s rank correlation coefficient between doses (1 = highest concentration; 7 = lowest concentration, 0 = no agonist) of agonists were calculated (n = 15 for each phase). All correlations (r > 0.5) are statistically significant (p < 0.001). Doses

1–2

2–3

3–4

4–5

5–6

6–7

7–0

ADP

0.92

0.91

0.89

0.82

0.83

0.94

0.89

U46619

0.88

0.88

0.92

0.92

0.80

0.85

0.84

TRAP

0.88

0.70

0.80

0.90

0.90

0.91

0.94

CRP

0.95

0.94

0.91

0.84

0.72

0.90

0.91

Table 3: Analysis of the variability of the data: good correlation of individual’s response to agonist. Ratios of activation were calculated as described in materials and methods for each donor and agonist. Spearman’s rank correlation coefficients between ratios of activation of agonists were calculated for all 15 donors. Statistically significant correlations (r > 0.5) are indicated in bold (p < 0.05). U46619

TRAP

CRP

Phase

1

2

1

2

1

2

ADP

0.63

0.38

0.70

0.86

0.65

0.84

0.71

0.59

0.86

0.61

0.80

0.92

U46619 TRAP


397

398


A

B

C

D

E

Figure 4: Healthy individuals display good concordance in platelet responses between phase 1 and 2 despite high inter-variability in EC50 values. A) Ratios of activation were calculated as described in Materials and methods and are presented as mean ± SE (n = 15). B) Box and whisker plots showing median and interquartile range of EC50 values (n = 15) calculated from non-linear regression curves fitted to binding of platelets

to SZ51 in the presence of increasing concentrations of ADP, U46619, TRAP or CRP. C) Representative graphs for each category of platelet response observed: low, normal and high responders to TRAP. EC50 values are indicated in each graph. Each data point represents mean of duplicate measurements ±SE.

(CD63) and phosphorylation of vasodilator-stimulated phosphoprotein (VASP) (16, 33, 34)] detectable by FACS-analysis, have been put forward as alternatives for aggregation testing; however, these measurements may still suffer from a lack of ‘high-throughput’ capacity. Advantages and disadvantages of platelet functional tests have been recently reviewed (14, 16, 35, 36), and although point-of-care assays are very attractive because they are easy to perform and to readout (16) they also lack sensitivity towards certain pathways e.g. those mediated through COX-1 and P2Y12 (36, 37). We here report a methodology that would overcome most of the above mentioned problems by using an ELISA setup in which activated platelets are captured by a coated antibody that specifically

distinguishes between resting and activated platelets, and next quantified by an antibody that is largely invariant to changes in the platelet activation state. Agonists (ADP, U46619, TRAP, CRP, AA) and antagonists (SNAP, PGE1, MRS-2179) of platelet activation could induce or prevent the binding of platelets to the capturing moAb in a dose-dependent manner, respectively. We illustrate the applicability of the new platelet ELISA for screening purposes including both platelet activation or inhibition pathways. In addition, whole blood can also be used successfully, indicating that much less blood is needed to carry out the platelet ELISA test, which constitutes a significant advantage over standard aggregometry or to a recently developed aggregation assay in a 96-well plate format that require 60 ml of blood to obtain washed platelet preparation (15). Additionally, in platelet disorders, such as BSS or giant platelet syndrome, where PRP can not readily be prepared this could provide a way to study their activation profiles in more detail. As expected, since the platelet assay is based on the detection of surface-exposed CD62P, the test has a low sensitivity to ASA, irrespective of whether ADP, U46619, TRAP or CRP is used as agonist. Indeed, incubation of PRP with high levels of ASA (100 μg/ml) did not influence the platelet response to any of these agonists. Similar results were also obtained by testing subjects (n = 4) one day post ASA intake (data not shown). These results were corroborated by data obtained in flow cytometry where CD62P expression in re-

Table 4: Analysis of the variability of the data in the assay: High interindividual variation. Coefficients of variation expressed in percentage between healthy subjects for each agonist and for each phase calculated for EC50 and ratios of activation.

Phase 1 EC50

Ratios

Phase 2 EC50

ADP

91.8

Ratios

69.6

57.6

56.0

U46619 TRAP

98.0

142.2

108.9

69.8

63.0

69.8

53.9

55.1

CRP

128.3

144.5

95.8

94.8


© Schattauer 2010


A

B

C

D

Figure 5: Platelet ELISA assay can be performed in whole blood. MoAb SZ51 (5 μg/ml) was coated in a 96-well microtiter plate overnight at 4°C and subsequently incubated with PRP (●) or whole blood (❍) (representative of four donors) in the presence of 0–20 μM ADP (A), 0–30 μM U46619

(B), 0–40 μM TRAP (C) or 0–125 ng/ml CRP (D). Platelet binding was detected by addition of biotinylated moAb 6B4 and streptavidin-HRP. Sigmoidal regression fit was applied to the curves and EC50 calculated using GraphPad software. All data points represent duplicate measurements (mean ± SE).

sponse to various concentrations of agonists in the presence of high concentrations of ASA (up to 1 mg/ml, data not shown) was not affected. Results obtained here are in accordance not only with flow cytometry studies showing no effect of ASA on the expression of P-selectin induced by weak or strong agonists on platelets from healthy subjects (28–30, 38, 39) but also with the low sensitivity of the detection of ASA effects in several point-of-care devices (e.g. PFA100, PlateletWorks) and other platelet function tests (36, 37, 40). It is generally accepted that ASA exerts an inhibitory effect on platelet function either for platelet aggregation or in flow cytometry when levels of P-selectin after AA stimulation are assessed (41, 42). In line with this, we could also demonstrate the inhibitory effect of ASA both in the platelet ELISA and by flow cytometry when platelets were stimulated with AA. This reiterates the importance of the assay choice when testing platelet sensitivity to ASA. As we also could detect an inhibitory effect in ADP-induced platelet activation when an inhibitor of P2Y1 was used in the platelet ELISA, it would be of interest to now to confirm this for an inhibitor of P2Y12 receptor and to test platelets from patients receiving different anti-platelet therapies (e.g. ASA vs. clopidogrel or prasugrel) to further assess the sensitivity of the platelet ELISA to these drugs.

A small study on 15 healthy individuals was carried out to evaluate the intra- and inter-variability of the test. Significant correlations were found between technical replicates, between adjacent dosages, and between ratios of activation for all agonists, suggesting that subjects demonstrated a conserved response to agonists as assessed using these parameters. Responses to the different platelet agonists were also evaluated by calculating EC50 values. A high variability was observed between individuals and for individuals on repeated testing. Variation in platelet responses to agonists between normal individuals has been reported in all platelet tests available, including flow cytometric assays (10, 43–45). In addition, it is known that many factors can influence platelet function including type of food intake, smoking, sport activity, and medication (46, 47). Although care has been taken to exclude individuals that took medication prior to blood sampling that could potentially affect platelet function, and to schedule appointments at comparable hours during both phases of testing, other non-controllable factors could influence the platelet response. Interestingly, when whole blood and PRP (prepared from the same sample) were processed in parallel in the ELISA, the EC50 values showed in fact much less variability. Although the intra-variability

© Schattauer 2010


399

400


What is known about this topic? ●

●

Platelet aggregometry is the golden standard for platelet functional testing. However, it remains time consuming and requires a fairly high sample volume. Point-of-care assays, although very attractive because of their simplicity and rapidity, have limited sensitivity for assessment of certain anti-platelet drugs.

What does this paper add? ●

●

●

We have developed a novel platelet functional assay that allows several different activation and inhibitory signalling pathways of platelet activation to be studied simultaneously. This test can be used with platelet-rich plasma as well as whole blood; therefore, only reasonable blood sample volumes are required. The test is based on the ELISA technique, could be automated and used with large numbers of subjects for screening purposes.

of the test complicated the task of correlating directly EC50 values from phase 1 and 2 of the study, categorisation of platelet response profiles in hypo-, normal- or hyper-responders showed good reproducibility. In another potential application, the use of an antibody against αIIbβ3, such as 16N7C2 or PAC-1 (recognising the activated form) to capture the platelets and an anti-GPIb as a detecting antibody, the more classical platelet dysfunction disorders such as GT and BBS would also readily be detected as indeed we could demonstrate for a GT patient. The data presented here are intended to provide proof-of-principle and at present do not demonstrate the clinical value of the test. Extended studies on platelets in healthy individuals, in patients with bleeding disorders and also in patients at risk for cardiovascular diseases are needed to be able to assess the full applicability of the test as a screening and diagnostic tool and re-address the intra- and inter-variability. The possibility of using whole blood instead of PRP, to pre-prepare the (ant)agonists at different concentrations, store the ELISA plates and to readily automate the test are potential advantages. To these may be added the simultaneous determination of a large number of platelet reactions and samples. We believe that with the present test we have devised a method that should provide a valuable addition to those currently available and with its high-throughput potential will be useful in both studies of large control and patient cohorts as required in functional genomic/proteomic approaches, as well as in screening tests for new antiplatelet agents. Acknowledgements

I.I.S. is a postdoctoral fellow and was supported by the European Union 6th Framework Program (LSHM-CT-2004–503485); K.B. is a postdoctoral fellow supported by a grant from the KU Leuven Industrial Research Fund (KP06/001). A.F. was an EU-RTN (HPRNCT-2002–00253) postdoctoral fellow; T.S. was supported by a BiThrombosis and Haemostasis 104.2/2010

lateral Collaboration between Flanders and Hungary (BIL/04/35) and the EU-RTN (HPRN-CT-2002–00253); K.V. is a postdoctoral fellow supported by the Fonds voor Wetenschappelijk Onderzoek (FWO), Vlaanderen. We would like to thank K. Freson (KU Leuven, Belgium) and K. Devreese (University Ghent, Belgium) for providing additional samples. We are grateful to F. Dudbridge (Institute of Public Health, Cambridge, UK) for his guidance and helpful discussions for statistical evaluation of the test. Finally, we thank all the donors that generously participated in this study.

References 1. Savage B, Almus-Jacobs F, Ruggeri ZM. Specific synergy of multiple substrate-receptor interactions in platelet thrombus formation under flow. Cell 1998; 94: 657–666. 2. Lecut C, Schoolmeester A, Kuijpers MJ, et al. Principal role of glycoprotein VI in alpha2beta1 and alphaIIbbeta3 activation during collagen-induced thrombus formation. Arterioscler Thromb Vasc Biol 2004; 24: 1727–1733. 3. Varga-Szabo D, Pleines I, Nieswandt B. Cell adhesion mechanisms in platelets. Arterioscler Thromb Vasc Biol 2008; 28: 403–412. 4. Jackson SP, Nesbitt WS, Kulkarni S. Signaling events underlying thrombus formation. J Thromb Haemost 2003; 1: 1602–1612. 5. Deckmyn H, Vanhoorelbeke K, Ulrichts H, et al. Amplification loops and signal transduction pathways. eds. Leuven University Press 2003; pp. 75–91. 6. Coughlin SR. Protease-activated receptors in hemostasis, thrombosis and vascular biology. J Thromb Haemost 2005; 3: 1800–1814. 7. Deckmyn H, Ulrichts H, Van De Walle G, et al. Platelet antigens and their function. Vox Sanguinis 2004; 87 (Suppl 2): 105–111. 8. Armstrong RA. Platelet prostanoid receptors. Pharmacol Ther 1996; 72: 171–191. 9. Mitchell JA, Ali F, Bailey L, et al. Role of nitric oxide and prostacyclin as vasoactive hormones released by the endothelium. Exp Physiol 2008; 93: 141–147. 10. Jones CI, Garner SF, Angenent W, et al. Mapping the platelet profile for functional genomic studies and demonstration of the effect size of the GP6 locus. J Thromb Haemost 2007; 5: 1756–1765. 11. Matzdorff A. Platelet function tests and flow cytometry to monitor antiplatelet therapy. Sem Thromb Hemost 2005; 31: 393–399. 12. Franchini M. The platelet-function analyzer (PFA-100) for evaluating primary hemostasis. Hematology 2005; 10: 177–181. 13. Zwierzina WD, Kunz F. A method of testing platelet aggregation in native whole blood. Thromb Res 1985; 38: 91–100. 14. Shah U, Ma AD. Tests of platelet function. Curr Opin Hematol 2007; 14: 432–437. 15. Moran N, Kiernan A, Dunne E, et al. Monitoring modulators of platelet aggregation in a microtiter plate assay. Analyt Biochem 2006; 357: 77–84. 16. Michelson AD. Methods for the measurement of platelet function. Am J Cardiol 2009; 103 (3 Suppl): 20A-26A. 17. Toth O, Calatzis A, Penz S, et al. Multiple electrode aggregometry: a new device to measure platelet aggregation in whole blood. Thromb Haemost 2006; 96: 781–788. 18. Peace AJ, Tedesco AF, Foley DP, et al. Dual antiplatelet therapy unmasks distinct platelet reactivity in patients with coronary artery disease. J Thromb Haemost 2008; 6: 2027–2034. 19. Cauwenberghs N, Vanhoorelbeke K, Vauterin S, et al. Epitope mapping of inhibitory antibodies against platelet glycoprotein Ibalpha reveals interaction between the leucine-rich repeat N-terminal and C-terminal flanking domains of glycoprotein Ibalpha. Blood 2001; 98: 652–660. 20. Wu G, Xi X, Li P, et al. Preparation of a monoclonal antibody, SZ-51, that recognizes an alpha-granule membrane protein (GMP-140) on the surface of activated human platelets. Nouv Rev Franc Hematol 1990; 32: 231–235. 21. Deckmyn H, Stanssens P, Hoet B, et al. An echistatin-like Arg-Gly-Asp (RGD)containing sequence in the heavy chain CDR3 of a murine monoclonal antibody that inhibits human platelet glycoprotein IIb/IIIa function. Br J Haematol 1994; 87: 562–571. 22. Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics 1989; 45: 255–268.

© Schattauer 2010


23. Stenberg PE, McEver RP, Shuman MA, et al. A platelet alpha-granule membrane protein (GMP-140) is expressed on the plasma membrane after activation. J Cell Biol 1985; 101: 880–886. 24. Chung AW, Jurasz P, Hollenberg MD, et al. Mechanisms of action of proteinaseactivated receptor agonists on human platelets. Br J Pharmacol 2002; 135: 1123–1132. 25. Cauwenberghs N, Meiring M, Vauterin S, et al. Antithrombotic effect of platelet glycoprotein Ib-blocking monoclonal antibody Fab fragments in nonhuman primates. Arterioscl Thromb Vasc Biol 2000; 20: 1347–1353. 26. Lopez JA. The platelet glycoprotein Ib-IX complex. Blood Coag Fibrinol 1994; 5: 97–119. 27. Milet-Marsal S, Breillat C, Peyruchaud O, et al. Analysis of the amino acid requirement for a normal alphaIIbbeta3 maturation at alphaIIbGlu324 commonly mutated in Glanzmann thrombasthenia. Thromb Haemost 2002; 88: 655–662. 28. Holmes MB, Sobel BE, Howard DB, Schneider DJ. Differences between activation thresholds for platelet P-selectin glycoprotein IIb-IIIa expression and their clinical implications. Thromb Res 1999; 95: 75–82. 29. Dropinski J, Jakiela B, Sanak M, et al. The additive antiplatelet action of clopidogrel in patients with coronary artery disease treated with aspirin. Thromb Haemost 2007; 98: 201–209. 30. Li N, Hu H, Hjemdahl P. Aspirin treatment does not attenuate platelet or leukocyte activation as monitored by whole blood flow cytometry. Thromb Res 2003; 111: 165–170. 31. Leon C, Ravanat C, Freund M, Cazenave JP, Gachet C. Differential involvement of the P2Y1 and P2Y12 receptors in platelet procoagulant activity. Arterioscler Thromb Vasc Biol 2003; 23: 1941–1947. 32. Born GV. Aggregation of blood platelets by adenosine diphosphate and its reversal. Nature 1962; 194: 927–929. 33. Metzelaar MJ, Korteweg J, Sixma JJ, et al. Comparison of platelet membrane markers for the detection of platelet activation in vitro and during platelet storage and cardiopulmonary bypass surgery. J Lab Clin Med 1993; 121: 579–587. 34. Pampuch A, Cerletti C, de Gaetano G. Comparison of VASP-phosphorylation assay to light-transmission aggregometry in assessing inhibition of the platelet ADP P2Y12 receptor. Thromb Haemost 2006; 96: 767–773. 35. Hayward CP, Eikelboom J. Platelet function testing: quality assurance. Semin Thromb Hemost 2007; 33: 273–282.

© Schattauer 2010

36. Gibbs NM. Point-of-care assessment of antiplatelet agents in the perioperative period: a review. Anaesth Intensive Care 2009; 37: 354–369. 37. Williams CD, Cherala G, Serebruany V. Application of platelet function testing to the bedside. Thromb Haemost 2010; 103: 29–33. 38. Chandler AB, Earhart AD, Speich HE, et al. Regulation of CD40L (CD154) and CD62P (p-selectin) Surface Expression upon GPIIb-IIIa Blockade of Platelets from Stable Coronary Artery Disease Patients. Thromb Res 2010; 125: 44–52. 39. Fox SC, May JA, Shah A, Neubert U, Heptinstall S. Measurement of platelet P-selectin for remote testing of platelet function during treatment with clopidogrel and/or aspirin. Platelets 2009; 20: 250–259. 40. Blais N, Pharand C, Lordkipanidze M, et al. Response to aspirin in healthy individuals. Cross-comparison of light transmission aggregometry, VerifyNow system, platelet count drop, thromboelastography (TEG) and urinary 11-dehydrothromboxane B(2). Thromb Haemost 2009; 102: 404–411. 41. McKenzie ME, Malinin AI, Bell CR, et al. Aspirin inhibits surface glycoprotein IIb/ IIIa, P-selectin, CD63, and CD107a receptor expression on human platelets. Blood Coagul Fibrinolysis 2003; 14: 249–253. 42. Hubl W, Assadian A, Lax J, et al. Assessing aspirin-induced attenuation of platelet reactivity by flow cytometry. Thromb Res 2007; 121: 135–143. 43. Panzer S, Hocker L, Koren D. Agonists-induced platelet activation varies considerably in healthy male individuals: studies by flow cytometry. Ann Hematol 2006; 85: 121–125. 44. Schmitz G, Rothe G, Ruf A, et al. European Working Group on Clinical Cell Analysis: Consensus protocol for the flow cytometric characterisation of platelet function. Thromb Haemost 1998; 79: 885–896. 45. Feijge MA, van Pampus EC, Lacabaratz-Porret C, et al. Inter-individual variability in Ca2+ signalling in platelets from healthy volunteers: effects of aspirin and relationship with expression of endomembrane Ca2+-ATPases. Br J Haematol 1998; 102: 850–859. 46. Walters J, Skene D, Hampton SM, Ferns GA. Biological rhythms, endothelial health and cardiovascular disease. Med Sci Monitor 2003; 9: RA1–8. 47. Peace AJ, Egan K, Kavanagh GF, et al. Reducing intra-individual variation in platelet aggregation: implications for platelet function testing. J Thromb Haemost 2009; 7: 1941–1943.


401