Correlations among different platelet aggregation pathways in a group of healthy volunteers

Abstract

Platelet aggregation is a complicated process mediated by different signaling pathways. As the process is highly complex and apparently redundant, the relationships between these pathways are not yet fully known. The aim of this project was to study the interconnections among seven different aggregation pathways in a group of 53 generally healthy volunteers aged 20 to 66 years. Platelet aggregation was induced with thrombin receptor activating peptide 6 (TRAP), arachidonic acid (AA), platelet activating factor 16 (PAF), ADP, collagen, thromboxane A₂ analogue U46619 or ristocetin (platelet agglutination) ex vivo in fasting blood samples according to standardized timetable protocol. Additionally, some samples were pre-treated with known clinically used antiplatelet drugs (vorapaxar, ticagrelor or acetylsalicylic acid (ASA)). Significant correlations among all used inducers were detected (Pearson correlation coefficients (r_P): 0.3 to 0.85). Of all the triggers, AA showed to be the best predictor of the response to other inducers with r_P ranging from 0.66 to 0.85. Interestingly, the antiplatelet response to ticagrelor strongly predicted the response to unrelated drug vorapaxar (r_P = 0.71). Our results indicate that a response to one inducer can predict the response for other triggers or even to an antiplatelet drug. These data are useful for future testing but should be also confirmed in patients.

Plain Language Summary

What is the context?

• Platelet activation is a complicated process with multiple signaling cascades involved.

• A total of seven common platelet triggers (ADP, collagen, TRAP-6, PAF, arachidonic acid/AA/, ristocetin and U46619) were tested.

• The process is dependent on many factors including sex, age, concomitant disease(s), pharmacotherapy.

What is new?

• There were significant correlations between all tested aggregatory cascades.

• AA has the highest rate of response predictability in our heterogeneous generally healthy volunteer group.

• There was no correlation between impedance aggregometry in whole blood and turbidimetric measurement with platelet-rich plasma.

What is the impact?

• The effect of antiplatelet drugs can be assessed from the reaction to different trigger(s) at least in this group of healthy patients.

• Future studies must test these relationships in patients with different diseases.

Graphical Abstract

Keywords:

• Antiplatelet drug
• arachidonic acid
• healthy volunteer
• platelet aggregation
• signaling pathway

Introduction

Blood hemostasis is a tightly regulated process in which complicated physiological pathways coordinate to avoid excessive blood loss upon endothelium rupture. The main players in this process are blood platelets, that are able to change their conformation after activation, secrete intracellular granules containing secondary mediators and associate with other activated platelets, forming a plug, which prevents bleeding.¹ Under normal circumstances, platelets are only activated upon endothelial injury, which leads to exposure of collagen from subendothelial layer. Collagen then interacts with platelets and von Willebrand factor (vWF) and this causes platelet adhesion and activation. The subsequent cascade is highly intricate and involves many receptors and signaling molecules.² All these cascades lead to the same final consequences: an increase in intracellular calcium, activation of glycoprotein IIb/IIIa receptor and clustering of platelets. Intracellular granule releases encompassing ADP, and thromboxane production seems to be a particularly crucial process, as they represent the targets of commonly used antiplatelet drugs.³ These drugs, comprehending cyclooxygenase inhibitors (ASA), ADP antagonists and some novel drugs, such as thrombin antagonist vorapaxar, are used for prevention of possible serious thrombosis mainly in secondary prevention of cardiovascular diseases of ischemic origin.^4,5

As different pathological conditions can induce or facilitate this activation even if the endothelial integrity has not been explicitly disrupted (e.g. atherosclerosis, diabetes mellitus, etc.), it is important to study different platelet activation pathways as these might be activated in cardiovascular diseases.² For triggering of certain aggregatory pathway, endogenous compounds can be used (e.g. ADP, arachidonic acid /AA/, platelet activating factor /PAF/ or collagen). However, some of these signaling molecules are unstable, too large or present multiple effects. Therefore, synthetic compounds mimicking platelet aggregation cascades are used (e.g. thromboxane analogue U-46619, vWF receptor binding activator ristocetin or thrombin sequence without enzymatic effect TRAP /thrombin receptor activating peptide/) in experimental conditions.

Most studies found in literature typically include one or two platelet triggers at most. However, platelet aggregation is a complicated process and the use of only some triggers may bias the results. We suppose that the use of more triggers can map the background of platelet aggregation process more accurately. Therefore, in this work we aimed to study the effect of seven relevant platelet activators (ADP, AA, collagen, PAF, TRAP, U-46619, and ristocetin) in blood samples of healthy patients (Figure 1). In addition, blood samples of healthy individuals were treated with these activators alone, or in combination with clinically used antiplatelet drugs (ASA, ticagrelor, and vorapaxar), to study the interrelationships between them more extensively.

Figure 1. Platelet activation pathways of the different platelet triggers and antiplatelet compounds used. Tested inducers are depicted as yellow ellipses while tested antiplatelet drugs as red ellipses. Abbreviations: AA: arachidonic acid; ADP: adenosine diphosphate; ASA: acetylsalicylic acid; COX-1: cyclooxygenase-1; cPLA2: cytosolic phospholipase A₂; DTS: dense tubular system; Gi: G protein α subunit i; GP Ib: glycoprotein Ib; GP VI: glycoprotein VI; Gq: G protein α subunit q; IP₃: inositol triphosphate, P2RY₁, P2RY₁₂: two receptors for ADP, PAF: platelet activating factor; PAFR: PAF receptor; PAR1: protease activated receptor 1; PIP₂: phosphatidylinositol 4,5-biphosphate; PKC: protein kinase C; TP: thromboxane receptor, TRAP: thrombin receptor-activating peptide; TXA₂: thromboxane A₂; α2β1: integrin α₂β₁.

Methods

Chemicals

Dimethyl sulfoxide (DMSO), ristocetin, PAF-16, ASA, ticagrelor and U-46619 were purchased from Sigma (St. Louis, MO, USA). AA, ADP, TRAP-6 were bought from Roche (Basel, Switzerland). Saline was purchased from B.Braun (Melsungen, Germany) while collagen was from Diagnostica, a.s. (Prague, Czech Republic).

Healthy donor enrollment

Details of the study were previously published.⁶ Shortly, blood samples from a total of 53 generally healthy human donors were collected by venipuncture into plastic disposable tubes containing heparin sodium (170 IU/10 mL). Additional 13 patients were enrolled to perform complementary experiments on optical aggregometer and to compare the reactivity by impedance aggregometer using citrated and heparinized blood. The sampling was always performed in the morning (7:30–8:30 a.m.). The inclusion criteria were age ≥20 years and subjective good health. Exclusion criteria were the presence of heart disease, diabetes mellitus of either type, peripheral artery disease, a history of stroke, and the use of drugs which directly affect platelet aggregation (e.g. antiplatelet drugs, non-steroidal anti-inflammatory drugs). Smoking, administration of oral contraception, and controlled hypertension were acceptable for enrollment in this study, to enrich the variability of the group. Alcohol was not allowed 24 hours prior to blood collection. All donors signed an informed consent in line with the approval of ethics committee of the University Hospital in Hradec Králové (No. 202007S01P from 18 June 2020). All experiments conformed to the latest Declaration of Helsinki.

Aggregation experiments

Blood was carefully transferred under standard conditions to our laboratory for analysis. Aggregatory experiments were initiated precisely 30 minutes from blood draw. Whole heparinized or citrated blood was used for all measurements with impedance aggregometer Multiplate (Roche Diagnostic, Basel, Switzerland). The method has already been described elsewhere.⁷ Shortly, 300 μL of whole blood was diluted with the same volume of preheated 0.9% sodium chloride, and platelet aggregation was induced with one of the following triggers: AA, collagen, PAF-16, ADP, U46619, TRAP-6, or ristocetin at desired concentration(s) and monitored for 6 min. For antiaggregatory experiments, a 3-minutes preincubation with the specified antiplatelet drugs at desired concentrations was performed, before triggering aggregation. A solvent (control) sample with DMSO at a final concentration of 0.8% was also included for each measurement.

Platelet-rich plasma (PRP) was prepared from citrated whole blood after centrifuging at 180 g for 10 min at room temperature (Eppendorf 5804 R, Hamburg, Germany). Platelet-poor plasma was prepared from the remaining blood at 2800 g (10 min, room temperature) and served as blank control for aggregation. These experiments were performed in optical aggregometer APACT 4S (Ahrensburg, Germany). Briefly, 200 μL of platelets set to a final concentration of 3 × 10⁸ platelets/mL was incubated for a minute before triggering aggregation with the same compounds used in impedance aggregation experiments. The reaction was followed for a total of 6 minutes, after which area under the curve (AUC), maximal aggregation, and velocity of the aggregation were recorded.

Statistical analysis

GraphPad Prism 9.3.1 (GraphPad Software, San Diego, CA, USA) was used for performing all data analyses. Correlations were tested by the Pearson´s correlation test. In case of significant correlations, linear regression was performed. Differences in aggregation were compared by paired t-test or the Wilcoxon tests depending on data distribution pattern.

Results and discussion

Aggregatory responses obtained in whole heparinized blood after treatment with different inducers documented significant relationships among all triggers used (Table I). Response to some of the used inducer significantly corresponded with a very high Pearson correlation coefficients (r_P) to reaction on another platelet aggregation inducer. This suggests biologically relevant relationships between all investigated aggregatory pathways. A relatively high number of individual samples (53 donors) accentuate the biological relevance of the observed finding. The strongest relationships were found for the higher concentration of AA (200 μM) with all triggers with the exception of the lower concentration of collagen 0.16 μg/mL (Figure 2a–c, Table I). Interestingly, the lower concentration of AA used (60 μM) also provided significant correlations with other platelet aggregation triggers, but weaker than those at 200 μM. Similar outcomes were observed with collagen, although these correlations were always weaker than those of AA. The strongest association obtained was for the higher concentrations of both AA and collagen (r_P = 0.85). These outcomes are understandable as collagen cascade is followed by a release of arachidonic acid with subsequent formation of thromboxane A₂.⁸ Collagen is one of the most studied platelet aggregation triggers⁹ as it initiates the process of platelet aggregation. It exerts its activity through interaction with different surface receptors GP VI and GP Ia/IIa (α2β1).¹⁰ Apparently, low concentration of collagen is only mildly proaggregatory and based on results from this study, higher concentrations collagen should be used in the experimentation.

Figure 2. The strongest associations between platelet aggregatory responses to inducers. (a): arachidonic acid (200 μM) and collagen (1 μg/mL), (d): arachidonic acid (200 μM) and U-46619 (80 nM), (c): arachidonic acid (200 μM) and TRAP (10 μM), (b): ADP (5 μM) and ristocetin (400 μM), (e): collagen (1 μg/mL) and acetylsalicylic acid (70 μM) + collagen (1 μg/mL), (f): ticagrelor 500 nM + ADP (5 μM) and vorapaxar (5 μM) and TRAP (10 μM). AUC: area under the curve measured by impedance aggregometry.

Table I. Relationships between responses to different aggregation inducers.

	AA 200	Collagen 0.16	Collagen 1	ADP	Ristocetin	TRAP	U-46619	PAF
AA60	0.47***	0.44**	0.45**	0.50***	0.46***	0.45**	0.55***	0.30*
	AA200	0.30*	0.85***	0.70***	0.70***	0.73***	0.75***	0.66***
		collagen 0.16	0.44**	0.43**	0.43**	0.32*	0.33*	0.39**
			collagen 1	0.59***	0.61***	0.72***	0.66***	0.55***
				ADP	0.81***	0.65***	0.56***	0.57***
					ristocetin	0.65***	0.61***	0.70***
						TRAP	0.53***	0.65***
							U-44619	0.66***

The data show the Pearson correlation coefficients of area under the curve for impedance aggregation measurements. *p < .05, **p < .01, ***p < .001. AA 60 – arachidonic acid 60 μM, AA 200 – arachidonic acid 200 μM, collagen 0.16 – collagen 0.16 μg/mL, collagen 1 - collagen 1 μg/mL, ADP −5 μM, ristocetin − 400 μM, TRAP (thrombin receptor activation protein 6) − 10 μM, U-44619 – 80 nM, PAF (platelet activating factor 16) − 20 nM. Bold values represent the highest correlations found (>0.7).

ADP, present in dense granules, is released after platelet stimulation with different inducers and further promotes platelet aggregation through interaction with platelet surface P2Y₁ and P2Y₁₂ receptors.¹¹ As vWF and collagen are involved in the first stages of the platelet aggregation process, significant correlations with ADP were expected.

Our results also showed a high correlation between ADP and AA. These pathways are, at first sight, unrelated. The observed correlation is likely consequence of AA-based thromboxane formation followed by stimulation of thromboxane receptors coupled with G_q, which stimulates phospholipase C. This leads to an increase in intracellular calcium concentration, which is a known trigger of platelet granule release encompassing ADP release. Similarly, thromboxane, thrombin and PAF activate receptors coupled with G_q. Strong correlations among platelet responses to these compounds with ADP were observed too. These results indicate that intracellular calcium increase with a subsequent release of ADP is at least partly mediating the response to thromboxane, thrombin, and PAF. The situation is more complicated with thromboxane and thrombin as their receptors couple both with Gq and G₁₃. The fact that AA is needed for synthesis of thromboxane can explain the strong correlation between AA and U-44619 (r_P = 0.75), a stable thromboxane A₂ analogue. The strongest correlation of ADP was that with ristocetin (r_P = 0.81; Figure 2d). Ristocetin, originally an antibiotic,¹² is employed as a useful pharmacological tool mediating activation of vWF leading to platelet agglutination. Ristocetin provided a higher correlation with ADP than with collagen. This might point out that vWF activation is linked with a more intense release of ADP than collagen itself.

PAF is a lipid-like molecule with complex effects including an ability as a potent platelet trigger.¹³ Its synthesis is a complicated process via two main pathways (de novo and remodeling) not only in platelets but also in leukocytes, endothelial and mast cells.¹⁴ PAF is able to activate platelets even at picomolar concentrations and has an important role in inflammation, recruiting phagocytes, leukocytes and macrophages. This molecule interacts with its specific receptor coupled with Gq on platelet surface. According to our results, it correlated significantly with all triggers while the highest correlation was obtained with ristocetin. This is a novel finding since no common signaling pathways are known for these triggers. Whereas ristocetin causes platelet agglutination through interaction with GP Ib, PAF has a different signalization pathway mediated by mentioned G_q protein. Correlations with thromboxane, TRAP and ADP pathways might be based on the fact, that all of them activate G_q-coupled receptors and hence have the same consequences, although there are some differences as described above.

TRAP is another synthetic molecule, acting directly on thrombin receptors PAR1 and PAR4.¹⁵ In our study, we chose this molecule instead of thrombin, as it does not activate blood coagulation, since it is free of the enzymatic activity but retains the ability to activate thrombin receptors on platelets. The strongest correlations were observed, quite surprisingly, with AA and collagen. The connection is not clear as there is not a direct connection among these signaling pathways. Regardless, it is possible that the found correlations are consequence of a higher platelet reactivity, which can be reflected in all aggregation pathways.

Since differences in platelet response to a given trigger can vary among the different aggregatory methods available, complementary experiments were performed. No correlation between platelet aggregation measured in whole citrated blood and PRP prepared from citrated blood by impedance and turbidimetric aggregometry, respectively, was found (Table S1). Only a nearly significant correlation was observed for AA (p = .055). There are two likely explanations: a) other blood cells can impact the process of platelet aggregation and/or b) PRP preparation affects platelet reactivity, e.g. by partly activating them. The latter can be supported by the fact that optical aggregation using PRP resulted in numerically lower platelet aggregation compared to whole blood impedance aggregation with the exception of ADP (Figure S1-S2). Another aspect investigated was the impact of used anticoagulants to drawn blood samples (heparin and sodium citrate). In impedance experiments, a significant correlation was observed solely for AUC in relation to collagen. Additionally, an almost significant correlation was observed for U-46619 (p = .051). This means, that solely collagen and partly U-46619 triggered platelet aggregation are giving results which are predictable by using citrate or heparinized blood. Moreover, when considering maximal aggregation values, no correlation was observed at all. Such results were unexpected. Heparin blocks coagulation via interaction with antithrombin III whereas citrate chelates free calcium. A previous study reported that both citrated and heparinized blood are not completely stable in time in relation to platelet count, therefore, this was likely not the reason for the observed differences in our results.¹⁶ For ADP, PAF and U-46619, no significant differences between platelet reactions in citrated and heparinized blood were observed. For the remaining triggers, the measured reaction, either as AUC and/or maximal aggregation, was stronger with citrated blood samples. It is known that heparin also interacts with von Willebrand factor, although the clinical relevance of this aspect is not clear.¹⁷ Regardless, AUC but not maximal aggregation after ristocetin was significantly lower in case of heparinized blood (Figure S1-S2). Whether this fact can explain the lower reaction to other inducers is not clear.

In the last step, the inhibitory capacity of antiplatelet drugs in heparinized blood was correlated with platelet aggregatory responses to the different triggers used. Observed correlations (Table II) can facilitate future research by assessing the response to particular drugs as it is not possible to test all aggregatory pathways in every patient. Inhibitor-trigger pairing with inducers were chosen according to their activation pathways and only clinically relevant drugs ASA (aspirin), ticagrelor and vorapaxar, were selected. The mechanism of action of ASA is well known, through irreversible blockade of platelet cyclooxygenase 1. The other two drugs target different cellular structures. Vorapaxar is a thrombin receptor antagonist (PAR1) whereas ticagrelor targets P2Y₁₂ receptor, preventing agonist binding.¹⁸ Our data provided significant correlations in the response to antiplatelet drug pre-treatment to most other inhibitors and inducers emphasizing a high degree of predictability even in unrelated cascades. An example of an expectable effect is the intensity of the platelet response to both low and high concentrations of collagen-induced aggregation, which was highly predictive to the effect of ASA 70 μM. Interestingly, the response to ticagrelor in ADP-triggered aggregation was also highly predictable to the response obtained after treatment with vorapaxar (5 µM) and TRAP (Figure 2f, Table I). This correlation was unexpected, as they are antagonists at two different receptors. Although one type of ADP receptor (P2Y₁) is G_q, like the TRAP receptor, this is not likely the explanation since ticagrelor selectively antagonizes the P2Y₁₂ (Gi) receptor for ADP. P2Y₁₂ antagonists have been reported to prevent thrmoboxane A₂ synthesis and this could explain the effect on other cascades.¹⁹

Table II. Relationships between responses to different platelet aggregation inhibitors and platelet aggregation inducers.

	AA 200 + ASA 30	AA 200 + ASA 70	collagen 1 + ASA 70	ticagrelor + ADP	vorapaxar 1 + TRAP	vorapaxar 5 + TRAP
AA 60	0.62***	0.45**	0.40**	0.35*	n.s.	0.40*
AA200	0.61***	n.s.	0.57***	0.56***	0.30*	0.37*
collagen 0.16	0.40**	0.63***	0.76***	n.s.	0.53***	0.59***
collagen 1	0.46***	n.s.	0.70***	0.59***	0.36**	0.37*
ADP	0.57***	n.s.	0.52***	0.59***	0.35*	0.46**
ristocetin	0.49***	0.32*	0.52***	0.47***	n.s.	0.48**
TRAP	0.49***	n.s.	0.55***	0.3*	0.47***	n.s.
U-44619	0.60***	n.s.	0.34*	0.43**	n.s.	0.37*
PAF	0.57***	0.36**	0.36**	n.s.	n.s.	n.s.
	AA 200 + ASA 30	0.53***	0.38**	0.43**	n.s.	0.42**
		AA 200 + ASA 70	0.40***	n.s.	n.s.	0.37*
			collagen 1 + ASA 70	0.43**	0.44**	0.34*
				ticagrelor + ADP	n.s.	0.71***

The data show the Pearson correlation coefficients of area under the curve for impedance aggregation measurements. *p < .05, **p < .01, *** p < .001. n.s.: non-significance. AA 60 – arachidonic acid 60 μM, AA 200 – arachidonic acid 200 μM, collagen 0.16 – collagen 0.16 μg/mL, collagen 1 - collagen 1 μg/mL, ADP − 5 μM, ristocetin −400 μM, TRAP (thrombin receptor activating protein 6) − 10 μM, U-44619 - 80 nM, PAF (platelet activating factor 16) − 20 nM, ASA 30 + AA 200 – acetylsalicylic acid 30 μM + arachidonic acid 200 μM, ASA 70 + AA 200 – acetylsalicylic acid 70 μM + arachidonic acid 200 μM, ASA 70 + collagen 1 - acetylsalicylic acid 70 μM + collagen 1 μg/mL, ticagrelor + ADP – ticagrelor 500 nM + ADP 5 μM, vorapaxar 1 + TRAP – vorapaxar 1 μM + TRAP 10 μM, vorapaxar 5 + TRAP – vorapaxar 5 μM + TRAP 10 μM. Bold values represent the highest correlations found (>0.7).

To our best knowledge, no previous scientific work so far has studied and compared the different activity of such a wide number of triggers. The main outcome is the fact that the magnitude of response to one platelet aggregation trigger can predict the response of another platelet signaling pathway as well as that of antiplatelet treatment. Our study has both strengths and limitations. The most important ones are: 1) data from 53 different volunteers were included, 2) the volunteers were of a wide age range, 3) on the other hand, a limitation of our study is the inclusion of healthy donors, 4) extrapolation of our data needs to be confirmed in cardiovascular patients

Availability of data and materials

Raw data from patients cannot be shared but are available upon reasonable request to the corresponding author.

Ethics approval

This study was performed in accordance to the Ethics Committee approval from the Faculty of Pharmacy, Charles University and the University Hospital Hradec Králové.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Czech Health Research Council [NU21J-02-00021], Charles University [SVV 260 549] and MH CZ - DRO [UHHK, 00179906].