Comparison of the GPVI inhibitors losartan and honokiol

Abstract

Losartan and honokiol are small molecules which have been described to inhibit aggregation of platelets by collagen. Losartan has been proposed to block clustering of GPVI but not to affect binding of collagen. Honokiol has been reported to bind directly to GPVI but only at a concentration that is three orders of magnitude higher than that needed for inhibition of aggregation. The mechanism of action of both inhibitors is so far unclear. In the present study, we confirm the inhibitory effects of both agents on platelet aggregation by collagen and show that both also block the aggregation induced by the activation of CLEC-2 or the low affinity immune receptor FcγRIIa at similar concentrations. For GPVI and CLEC-2, this inhibition is associated with a reduction in protein tyrosine phosphorylation of multiple proteins including Syk. In contrast, on a collagen surface, spreading of platelets and clustering of GPVI (measured by single molecule localisation microscopy) was not altered by losartan or honokiol. Furthermore, in flow whole-blood, both inhibitors suppressed the formation of multi-layered platelet thrombi at arteriolar shear rates at concentrations that hardly affect collagen-induced platelet aggregation in platelet rich plasma. Together, these results demonstrate that losartan and honokiol have multiple effects on platelets which should be considered in the use of these compounds as anti-platelet agents.

Keywords: CLEC-2; GPVI; honokiol; losartan; platelets.

Figure 1.

Losartan and honokiol dose-dependently reduce aggregation of washed platelets induced by collagen or rhodocytin. Washed platelets were pretreated with different concentrations of losartan or honokiol for 3 min before stimulation with a. collagen 1µg/ml or b. rhodocytin 150 nM. Bar graphs represent results from a total of 5 independent experiments and results are shown as mean ± SD **P < .01, ***P < .001.

Figure 2.

Losartan and honokiol inhibit platelet protein tyrosine phosphorylation, including Syk, induced by collagen or rhodocytin. Washed platelets were pre-treated with 25 µM of losartan or honokiol for 3 min before stimulation with collagen 3 µg/ml or rhodocytin 150 nM. Stimulations were stopped with addition of lysis buffer. a. whole-cell lysate (WCL) and b. immunoprecipitate (IP) of Syk. WCL and IP were separated by SDS-polyacrylamide gel electrophoresis and Western blotted for pTyr and Syk, respectively. The first lane shows the molecular weight markers (MW).The results are representative of 3 independent experiments.

Figure 3.

Losartan reduces signalling induced by collagen in transfected DT40 cells. DT40 B-cells were transfected with a NFAT-luciferase reporter construct, a β-galactosidase construct to control for transfection efficiency, and two GPVI and FcRγ-chain(Fcγ) expression constructs or empty control (CTL) vectors. Sixteen hours post-transfection, expression of GPVI was confirmed by flow cytometry (data not shown). The transfected cells were pre-treated with losartan (LOS) at 25 μM or 250 μM. They were either a. stimulated with collagen (10 μg/mL) or PMA (50 ng/mL) plus ionomycin (1 μM) or b. unstimulated. Six hours later, cells were lysed and assayed for luciferase and β-galactosidase. Luciferase data were normalized for β-galactosidase expression. Error bars represent the SEM from three independent experiments. * P < .001

Figure 4.

Losartan and honokiol do not affect platelet spreading on collagen. Glass coverslips were coated with collagen.Washed platelets were pre-treated with 25 µM of a. losartan or b. honokiol before spreading on collagen, followed by fixation and staining for actin with Alexa-488 phalloidin. Scale bar, 5 μm. Bar graphs illustrate quantification of the surface area covered with platelets per field, and the numbers of platelets counted per field. Dataare shown as mean ± SD and are representative of three experiments.

Figure 5.

Losartan and honokiol do not affect clustering of GPVI. Glass coverslips were coated with CRP. Washed platelets were pre-incubated with 1G5 Fab (pan-GPVI), then pre-treated with 25 µM losartan, 25 µM honokiol or vehicle before spreading on CRP. Following fixation, GPVI was secondary labelled with mouse-Alexa647 and actin was labelled with Alexa488-phalloidin. The single molecule localisation data acquired by STORM was analysed for clustering using DBSCAN. Different colours represent different clusters (with black dots representing noise).For each condition, 5 different fields of view (FOV) from 3 independent experiments were captured. Representative cropped scatter plots for each condition show the output of the DBSCAN clustering algorithm (left panel). Scale bars: 1 µm and 0.1 µm. Bar graphs show quantification of cluster density, cluster area and number of detections per cluster. Mean ± SEM (n = 3) (right panel).

Figure 6.

Losartan inhibits the formation of contracted, multilayered platelet thrombi under flow. Whole blood treated with vehicle solution or losartan was perfused over collagen surface for 6 min at wall shear rate of 1000 s⁻¹. Phase contrast images were captured and shown as representative images a. Platelet surface area coverage (% SAC), thrombus morphological score (range 0–5), thrombus contraction score (0–3) and platelet multilayer score (0–3) were determined b-e. Data are presented as mean ± SEM (n = 4). *P < .05, **P < .01, ***P < .001.

Figure 7.

Honokiol inhibits both platelet adhesion and formation of multilayered platelet thrombi under flow. Whole blood treated with vehicle solution or honokiol was perfused over collagen type I for 6 minutes at wall shear rate of 1000 s⁻¹. Phase contrast images were captured and shown as representative images a. Platelet surface area coverage (% SAC), thrombus morphological score (range 0–5), thrombus contraction score (0–3) and platelet multilayer score (0–3) were determined b-e. Data are presented as mean ± SEM (n = 4). *P < .05, **P < .01, ***P < .001.