Blood collection and the preparation of platelet suspensions and reconstituted blood

The local Medical Ethical Committee of University Medical Center Utrecht approved the drawing of blood for ex vivo research purposes, including those of this study. All donors had provided written informed consent and remained anonymous throughout the study. Blood was drawn by venipuncture from the antecubital vein of healthy donors or self-reported Cannabis sativa consumers [19] and collected into Vacuette tubes (Greiner, Kremsmünster, Austria) with 3.2% (m/v) trisodium citrate. In our study, we defined persons that self-reportedly used Cannabis sativa >4 times per week as cannabis consumers. No medication use was allowed. For controls, we included donors that had self-reportedly not consumed Cannabis sativa or other drug substances for at least 10 days before the experiments. Platelets were isolated as previously described. Reconstituted blood was prepared with washed platelets resuspended in 4 mM KCl, 124 mM NaCl, 20 mM NaHCO3, 2 mM Na2SO4, 1.5 mM MgCl2, 5 mM CaCl2, 5 mM D-glucose, 4% HSA at pH 7.3 to a platelet count of 3.33×1011platelets/L. Red blood cells were isolated by centrifugation at 2,000×g for 10 minutes. The supernatant plasma was removed and the pellet was resuspended, washed twice by centrifugation for 5 minutes at 1,700×g and packed by centrifuging at 2,200×g for 12 minutes in 0.9% NaCl and 5 mM D-glucose. Red blood cells were added to the washed platelet suspension to obtain a 40% hematocrit with a platelet count of 2×1011 platelets/L. All cell counts were routinely determined on a CellDyn 1800 (Abbott, Hoofddorp, The Netherlands).

Platelet aggregometry

Platelet-rich plasma (PRP) or washed platelets that had been pretreated with anandamide or vehicle (for 40 minutes unless indicated otherwise) were monitored in an optical aggregometer (Chrono-Log Corporation, Haverford, PA) at 37°C with a stirring speed of 900 revolutions per minute. Changes in light transmission in the platelet suspension were recorded for 20 minutes.

Flow cytometric determination of platelet P-selectin expression and glycoprotein IIb/IIIa activation

Platelet responsiveness to a concentration series of CRP-XL or TRAP was determined through binding of PE-conjugated mouse anti-human P-selectin IgG (125 in-assay dilution; BD Pharmingen, San Diego, CA) and 30 µg/mL Alexa-488-labeled fibrinogen (Invitrogen, San Diego, Ca). The platelet activation assay was carried out in a total volume of 55 µL by addition of 5 µL washed platelets, pretreated with anandamide or vehicle for 40 minutes. Next, the platelets were stimulated with CRP-XL, TRAP or U46619 at the indicated concentrations for 20 minutes. Finally, the samples were fixed with 500 µL 0.2% formyl saline (0.2% formaldehyde, 0,9% NaCl) and kept at room temperature until analyses on a FACSCalibur flow cytometer (Beckton Dickinson, Breda, the Netherlands).

Platelet aggregate formation under flow

Perfusions were carried out in a single-pass perfusion chamber according to established methods that were developed in-house [20]. In brief, 24×50 mm glass coverslips were coated with 100 µg/mL type I collagen (Horm; Nycomed, Linz, Austria) or 100 µg/mL human fibrinogen for 90 minutes at RT and subsequently blocked overnight with 1% HSA at 4°C. Reconstituted blood or whole blood was perfused over collagen-coated coverslips at a shear rate of 1600 s−1for 5 minutes. Washed platelets were perfused over fibrinogen-coated coverslips for 20 minutes at a constant flow rate of 25 s−1. Whole blood perfusion experiments were performed in the presence of 50 mM PPACK and 4300 U/mL pentasaccharide. Platelet adhesion and aggregate formation were continuously monitored using differential interference contrast microscopy with a Carl Zeiss AxioObserver Z1 and recorded through AxioCam MRm using Carl Zeiss AxioVision imaging software (Carl Zeiss MicroImaging GmbH, Gottingen, Germany) with a frequency of 1 frame/sec. A Zeiss 100x/1.3 EC PlanNeoFluar oil immersion lens was used for analysis on fibrinogen surfaces and Zeiss 40x/0.75 M27 for collagen surfaces.

Ca2+ mobilization

PRP was incubated with 3 µM Fura 2-AM (45 minutes, 37°C, light-protected). After incubation, PRP was acidified with ACD to pH 6.5, centrifuged again (330×g, 15 minutes, 20°C), and resuspended in HEPES/Tyrode buffer (145 mM NaCl, 5 mM KCl, 0.5 mM Na2HPO4, 1 mM MgSO4, 10 mM HEPES, pH 7.25) containing 5 mM D-glucose. The final platelet concentration was adjusted to 2.0×1011 platelets/L. Fura-2 fluorescence was recorded in 1.0-mL aliquots of platelet suspension without additional Ca2+ at 20°C in a F-7000 fluorescence spectrophotometer (Hitachi Ltd., Tokyo, Japan) with excitation wavelengths of 340 and 380 nm and emission at 510 nm. Changes in cytosolic [Ca2+] levels were monitored using the Fura-2 fluorescence ratio and calibrated on non-activated control samples, as well as EDTA (Fmin) and digitonin-treated (Fmax) samples for reference purposes. To investigate the influence of anandamide on calcium mobilization, platelets were pre-incubated at 37°C for 40 minutes with 50 µM anandamide and subsequently stimulated by collagen, CRP-XL, ADP, U46619, TRAP or human α-thrombin during measurements.

Immunoprecipitation and western blotting

Platelets were stirred at 900 rpm and pre- incubated at 37°C for 40 minutes with and without 50 µM anandamide and subsequently stimulated with either 1 µg/mL collagen or 0.1 U/mL human α-thrombin for 5 or 10 minutes. Platelet suspensions (450 µL of 2×1011platelets/L) were lysed in RIPA lysis buffer (1% m/v Nonidet P-40, 0.5% m/v Octylglucoside, 0.1% SDS, 5 mM EDTA), supplemented with 1 mM orthovanadate for 15 minutes on ice to prevent dephosphorylation after lysis. Lysates were mixed with 55 µL (10% vol/vol) protein G sepharose that had been pre-adsorbed with 1 mg/mL anti–Syk IgG (sc573; Santa Cruz Biotechnology, Santa Cruz, CA,) for 60 minutes at 4°C under rotation. Proteins were separated by sodium dodecylsulfate-polyacrylamide gel electrophoresis and subjected to western blotting. After incubating with blocking buffer (Odyssey; LI-COR, Lincoln, NE), 1 mg/mL anti-Syk IgG or anti-phosphotyrosine IgG (4G10; Upstate Biotechnology, Lake Placid, NY) was added to the membranes and finally the protein bands were visualized with an Odyssey Imaging system (LI-COR Biosciences, Lincoln, NE). Quantification was performed with Image-J software (NIH, Bethesda, MD). Possible lane-to-lane loading variation was corrected by normalization to the total amount immuno-precipitated Syk protein.

Anandamide inhibits α-granule secretion and limits platelet aggregate formation

In order to elucidate what mechanisms are involved in the inhibition of platelet aggregation by anandamide, we first examined α-granule release during platelet activation by determining P-selectin externalization by flow cytometry. When platelets are stimulated with either 2.5 µg/mL CRP-XL (Fig. 2A), 2.5 µM U46619 (Fig. 2B) or 5 uM TRAP (Fig. 2C), all platelets externalize P-selectin. In comparison, unstimulated platelets retain P-selectin intracellularly (not shown). Anandamide inhibits P-selectin expression by platelets in a dose-dependent fashion when they are stimulated with CRP-XL, U46619, or TRAP.

Figure 2

Anandamide inhibits platelet α-granule secretion and limits platelet aggregate formation.

We next analyzed whether the exposure of platelets to anandamide also affects aggregate buildup under flow. When untreated reconstituted blood is perfused over collagen at an arterial shear rate (1600 s−1), the adhesion of untreated platelets and subsequent aggregation occurs within 5 minutes (Fig. 2D, indicated as “control”:). In contrast, anandamide-treated platelets adhere to the collagen surface but are unable to form normal aggregates. In line with our earlier aggregometry experiments (Fig. 1), we found that this inhibitory effect of anandamide is dose-dependent (Fig. 2D). At lower anandamide concentrations (10 and 20 µM), platelet aggregate formation is comparable to that of the untreated platelets. However, at higher doses (30–50 µM anandamide), platelet aggregation becomes perturbed to an equal extent. We observed little difference at anandamide concentrations over 20 µM, which may be caused to the disaggregating effect of shear flow, which was not present in our earlier aggregometry studies.

To study the platelet-platelet interactions and aggregate stability, we subsequently perfused buffer for one minute over preformed aggregates. In contrast to aggregates of untreated platelets, this extra washing step further destabilized platelet clots that previously received high doses of anandamide and induced the formation of a disperse monolayer of discoid-shaped platelets on the collagen surface (Fig. 2E; the images in the right panel show platelets that were preincubated with 50 µM anandamide). These experiments indicate that anandamide modulates platelet reactivity by limiting α-granule secretion and reducing platelet aggregate formation under flow.

Anandamide inhibits platelet spreading on fibrinogen by limiting glycoprotein IIb/IIIa activation

We next investigated the effects of anandamide on the activation of glycoprotein IIb/IIIa (GPIIb/IIIa) on platelets by determining the amount of platelet-bound FITC-labeled fibrinogen by FACS. We found that anandamide inhibits GPIIb/IIIa activation, when triggered by 2.5 µg/mL CRP-XL (Fig. 3A) or 5 µM U46619 (Fig. 3B). However, only high concentrations (50 µM) of anandamide were able to reduce TRAP-induced fibrinogen binding activation (Fig. 3C). These data indicate that anandamide interferes with the fibrinogen-binding capacity of platelets, but apparently not sufficiently to inhibit platelet aggregation in our earlier aggregometry experiments (Fig. 1). These findings are in agreement with previous studies that reported that TRAP-induced platelet aggregation can occur partially independently of GPIIb/IIIa [21], [22].

Figure 3

Anandamide reduces glycoprotein IIb/IIIa activation and inhibits platelet spreading.

The adhesion and spreading of platelets on immobilized fibrinogen at a low shear rate (25 s−1) is dependent on GPIIb/IIIa. We next investigated the effects of anandamide on this process. While control platelets adhere to fibrinogen and gradually spread to cover the surface (Fig. 3D; quantification of the spreading behaviour of individual platelets in Fig. 3E), anandamide-treated platelets adhere to fibrinogen in comparable quantities (quantified in Fig. 3F) but are unable to spread normally (Fig. 3D, E).

Anandamide blocks collagen-induced platelet signaling

We earlier found that anandamide inhibits platelet aggregation induced by collagen and CRP-XL (Fig. 1), which requires signaling events through GPVI and Ca2+ mobilization. Phosphorylation of spleen tyrosine kinase (Syk) by Src-family kinases represents a key intermediate step in this pathway. The intracellular events involved in the suppressing effect of anandamide were first studied by investigating the intracellular Ca2+ levels in platelets after stimulation with collagen, CRP-XL, as well as with ADP, U46619, TRAP and thrombin (Fig. 4A). We found that the mobilization of Ca2+ in platelets that are activated by collagen and CRP-XL is significantly reduced in the presence of anandamide (Fig. 4A), while Ca2+mobilization remained unaffected when platelets are activated by ADP, U46619, TRAP or thrombin (Fig. 4A). Moreover, when analyzing the phosphorylation status of Syk in collagen-stimulated platelets, the characteristic time-dependent increase in phosphorylation was absent in the presence of anandamide, as well as in the presence of control inhibitor PP2 (a Src-family kinase inhibitor; representative western blot shown in Fig. 4B and densitometric quantification of repeated experiments in Fig. 4C). Our results show that anandamide inhibits Ca2+ mobilization and prevents Syk phosphorylation in platelets that are stimulated by collagen, which suggest that anandamide has the capacity to block specific elements of the collagen pathway.

Figure 4

Anandamide inhibits glycoprotein VI-dependent calcium mobilization and Syk-phosphorylation.

Platelet inhibition by anandamide is not dependent on platelet preactivation

The biological effects of anandamide in vivo are tightly regulated through degradation and uptake mechanisms that limit its extracellular lifetime [23]. In platelets this is amongst others mediated by fatty acid amide hydrolase (FAAH), which hydrolyzes anandamide into ethanolamine and AA. This latter molecule is a precursor of thromboxane A2 (TXA2), which is a well-known platelet agonist. Although our earlier aggregometry experiments suggested otherwise (Fig. 1G), the observed lowered platelet activation in the presence of anandamide could potentially be an indirect result of desensitization through TXA2 formation. URB597 is a selective inhibitor of FAAH that can be applied in vivo to prevent anandamide degradation, resulting in elevated tissue levels of anandamide and enhanced biological effects [24]. To explore the possibility that anandamide required processing by FAAH to inhibit platelet activation, we investigated the effects of URB597 on platelet inhibition by anandamide. After 20 minutes of pre-incubation with 25 µM URB597 to block FAAH, platelets were incubated with an intermediate concentration of 10 µM anandamide for an additional 40 minutes. We found that the inhibitory properties of anandamide are not negatively influenced by URB597 (Fig. 5A–F), indicating that FAAH activity is not required for platelet inhibition by anandamide. As expected, platelet activation by TRAP is unaffected by anandamide, and remains unaffected in the presence of URB597 (Fig. 5F). Furthermore, we found that URB597 significantly enhances the effects of anandamide when platelets are activated by collagen (Fig. 5A). Finally, we found that URB597 independently inhibited U46619-induced platelet aggregation (Fig. 5E), which could point to a potential contribution of FAAH to platelet activation. These experiments show that the inhibitory capacity of anandamide either slightly increases (collagen) or remains unchanged (CRP-XL, AA, ADP, U46619) when FAAH is blocked, arguing against a role for anandamide-induced platelet preactivation of platelets.

Figure 5

The inhibitory effect of anandamide on platelet activation is not dependent on platelet preactivation.

In a following experiment, we pretreated platelets with indomethacin and investigated the inhibitory potential of anandamide on collagen stimulation. Indomethacin blocks the conversion of arachidonic acid into TXA2, reducing platelet excitability. Accordingly, if platelet inhibition by anandamide is the result of preactivation (through the release of AA), indomethacin would reverse it. However, we found that the inhibitory effect of anandamide is persistent in the presence of indomethacin (Fig. 5G), but both inhibitors are not synergistic together. Together, these data show that the inhibitory effect of anandamide on platelets is independent of AA formation.

The authors thank Arjan Barendrecht, Steven de Maat, Tobias Nespital, Bert Rutten, Remo Diego Strangi for their technical support and Attie Tuinenburg for valuable discussion. We thank Prof. R. Farndale for kindly donating CRP-XL.