Abdu I. Alayash, PhD, Laboratory of Plasma Derivatives, CBER, FDA, Building 29, Room 112, 8800 Rockville Pike, Bethesda, MD 20892, USA. E-mail: firstname.lastname@example.org
Several lines of evidence point to the potential role of nitric oxide (NO) in the pathophysiology, as well as in the therapy, of sickle cell disease (SCD). In this study, we compared the effects of NO on platelets from normal individuals and from patients with SCD. Three NO donors were used to deliver NO to platelets: sodium 2-(N, N-diethylamino)-diazenolate-2-oxide (DEANO), S-nitrosocysteine (CysNO) and sodium trioxdintrate (OXINO or Angeli's salt). ADP-induced platelet aggregation, CD62P expression, PAC-1 binding and calcium elevation were evaluated in paired studies of normal and SCD subjects. DEANO significantly reduced aggregation in SCD platelets compared with normal platelets. DEANO similarly reduced the extent of CD62P expression in SCD platelets. All NO donors reduced PAC-1 binding, but there were no significant differences between platelets from normal or SCD subjects. Calcium elevation, as induced by ADP, was not altered by the presence of NO donors. However, when platelets were stimulated with thrombin, there was an increased initial response of SCD platelets compared with normal platelets. Taken together, these data suggest that the mode of NO delivery to platelets may produce various physiological responses and the optimization of NO delivery may contribute to reducing platelet aggregation in sickle cell disease.
The primary pathogenic defect in sickle cell disease (SCD) is believed to be the abnormal genetic disposition of sickle cell haemoglobin to polymerize under hypoxic conditions. Nitric oxide (NO), a key vasoactive molecule of the vascular system, has emerged as an important signalling molecule of the vascular system. NO inhibits adherence of white blood cells, red cells and platelets to endothelium (Radomski et al, 1987; Space et al, 2000). High levels of NO metabolites were found in the blood of SCD patients and were correlated with lower clinical pain scores during painful crises, suggesting additional potential beneficial actions of NO in SCD (Rees et al, 1995; Lopez et al, 1996). Recent clinical studies in patients with SCD reported that NO can augment blood oxygen affinity and may potentially ameliorate the severity of this disease (Head et al, 1997).
We have systematically investigated the potential therapeutic value of NO in SCD and first measured blood oxygenation in both normal and SCD patients who were given low-dose NO inhalation treatment (80 p.p.m.) for 2 h. We found no changes in the oxygen affinity of blood in both groups, in contrast to a previous study (Head et al, 1997), but elevated levels of methaemoglobin were observed (Gladwin et al, 1999a). We also reported similar in vitro effects when NO donor compounds were used to deliver NO to normal and sickle erythrocytes (Hrinczenko et al, 2000a). The underlying kinetics and the mode of NO transport across the erythrocyte membrane is currently under investigation by our group (Hrinczenko et al, 2000b).
Increased platelet activation in patients with SCD has been reported; however, very little is known regarding the precise mechanism (Wun et al, 1998). In the present study, we examined the effects of NO produced by NO donor compounds such as sodium 2-(N, N-diethylamino)-diazenolate-2-oxide (DEANO), S-nitrosocysteine (CysNO) and sodium trioxdintrate (OXINO or Angeli's salt). We measured platelet aggregation, CD62P expression, PAC-1 binding and calcium elevation in response to ADP stimulation in paired studies of normal and SCD subjects in the presence of these donor compounds. Among the donors used in this study, DEANO had the greatest effect upon ADP-induced platelet aggregation and CD62P expression. PAC-1 binding was reduced in the presence of all three NO compounds, however, there was no difference between the two platelet populations. Calcium elevation, as induced by ADP, was completely unaffected by any of the NO donors. The interplay between these three different NO donor compounds and platelets from SCD patients and its implication for the understanding of the role of NO in the pathophysiology of SCD is discussed.
Patients and methods
Sample preparation All subjects signed an informed consent document from a protocol approved by the National Heart, Lung and Blood Institute's Institutional Review Board. Blood was drawn from the antecubital vein of clinically stable SCD patients. Haemoglobin electrophoresis was performed to confirm haemoglobin S or A phenotype, as well as haemoglobin F levels. The SCD patients had not received hydroxyurea or butyrate therapy in the past 12 months, were non-smokers and had not received a blood transfusion in the past 3 months. Blood was drawn into the anticoagulant buffered citrate (129 mmol/l sodium citrate and 40 mmol/l citric acid, pH 5·6) at a ratio of 9:1. Platelet-rich plasma (PRP) was obtained by centrifugation at 135 g for 20 min. The platelet count was adjusted to 250 × 109/l with autologous platelet-poor plasma (PPP) obtained by centrifugation of residual red cells at 1500 g for 20 min. Control platelets were isolated using the same manner from normal subjects who had not taken any aspirin during the previous week. All experiments were performed on a minimum of four normal and four patient subjects.
NO donors DEANO (sodium 2-(N, N-diethylamino)-diazenolate-2-oxide) was obtained from Alexis Biochemicals (San Diego, CA, USA). OXINO or Angeli's salt (sodium trioxodinitrate) was the generous gift of Dr David A. Wink of the Radiation Biology Branch, National Cancer Institute, NIH, USA. To ensure that these compounds did not release NO until they were added to the platelets (which were at a pH of 7·4), they were diluted with 0·01 mol/l NaOH and kept on ice in the dark until ready to be added to the platelets. CysNO (S-nitrosocysteine) was prepared using standard methods according to Ferranti et al (1997). Briefly, equal parts of solution A (100 mmol/l l-cysteine and 0·1 mmol/l Na2EDTA in 250 mmol/l HCl) and solution B (100 mmol/l NaNO2 in H2O) were combined to yield a 50 mmol/l stock solution. CysNO is stable at a low pH, thus 0·01 mol/l HCl was used as the diluting agent; this compound remained on ice in the absence of light until added to the platelets. All NO donors were added to platelets at a final concentration of 20 μmol/l; this concentration was chosen because of its effects on sickle haemoglobin (Hrinczenko et al, 2000a). The volume of each NO donor required to attain the desired concentration was so small (1 μl in 250 μl of PRP) that it did not affect the pH of the platelet/plasma solution.
Platelet aggregation Aggregation was performed using a BioData PAP-4 aggregometer (Horsham, PA, USA). Platelet-rich plasma (PRP) was diluted to 250 × 109 platelets/l using platelet-poor plasma (PPP). Platelets were preincubated with the NO donor for 2 min at 37°C before addition of the agonist. Aggregation was induced using 10 μmol/l ADP and is reported as the peak value ± standard error of the mean (SEM). Statistical comparisons in all experiments were performed using Student's t-test and results were considered statistically significant when P < 0·05.
Flow cytometry PRP was diluted to 30 × 109/l with Walsh buffer, which is a modified Tyrode's buffer containing 140 mmol/l NaCl, 2·7 mmol/l KCl, 0·4 mmol/l NaHCO3, 0·1% glucose, 0·2 mmol/l MgCl2 and 0·1% fatty acid-free bovine serum albumin (Shattil et al, 1987). For CD62P expression, the platelets were incubated with a phycoerythrin (PE)-conjugated antibody against CD62P (Becton Dickinson, San Jose, CA, USA) and stimulated with 10 μmol/l ADP for 30 min at room temperature. Antibody binding was analysed using a FACScan flow cytometer (Becton Dickinson). Platelets were identified by characteristic light scatter, i.e. measuring forward and side scatter on a logarithmic scale. Results are expressed as the percentage of positive cells and were calculated by gating outside the isotypic control. PE-conjugated IgG1 (Becton Dickinson) was used as an isotypic control. PAC-1 is a monoclonal antibody (IgM) that preferentially binds the activated conformation of the platelet fibrinogen receptor, glycoprotein (GP)IIb–IIIa (Shattil et al, 1985). For PAC-1 binding experiments, platelets were incubated with fluorescein isothiocyanate (FITC)-conjugated PAC-1 (Becton Dickinson) and activated with 10 μmol/l ADP. Specific binding was defined by gating outside the isotype control, FITC-IgM (Caltag, Burlingame, CA, USA). Calcium elevation was measured by flow cytometry using platelets loaded with the fluorescent calcium indicator Fluo-4 (Molecular Probes, Eugene, OR, USA). Fluorescence was measured over a period of 3·5 min using the time acquisition mode of the flow cytometer with thrombin (positive control) present at a concentration of 1 U/ml; the negative control was the addition of 6 mmol/l EGTA. In all experiments, NO donors were present at a final concentration of 20 μmol/l and were incubated with the platelets for 2 min before stimulation with 10 μmol/l ADP. Aggregation experiments were performed before each calcium study to determine the concentration of ADP at which there was the greatest difference in aggregation between the patient platelets and the normal platelets. This concentration ranged from 0·5–2 μmol/l. Data was analysed using cellquest software (Becton Dickinson, version 3.2.1).
Figure 1 shows the plot of mean values ± SEM of peak aggregation induced by 10 μmol/l ADP or ADP plus NO donor and spontaneous aggregation. There was no significant difference in ADP-induced aggregation between normal platelets and SCD platelets. DEANO significantly reduced platelet aggregation in platelets from SCD patients, reducing the peak of ADP-induced aggregation from 75% to approximately 30%. OXINO slightly decreased aggregation in both platelet populations from 75% to approximately 55% and CysNO further decreased the response down to 40%, also equally in platelets from normal and SCD subjects. Figure 2 contains representative aggregation tracings from a normal (A) and a SCD subject (B). Over time, all the NO donors reduced the amount of platelet aggregation, but DEANO is exceptional in that it significantly reduced the maximum aggregation response in SCD platelets. CysNO dramatically inhibited the aggregation response of platelets from the SCD patient in this experiment. However, we saw a large variation in the responses of all platelets to this NO donor, both in aggregation and flow cytometry studies. In aggregation studies, the response of normal platelets to ADP and CysNO ranged from 15–46% aggregation; these values ranged from 3–35% aggregation in platelets from SCD patients.
Figure 3 shows the plot of mean values ± SEM of CD62P expression induced by ADP in the presence or absence of NO donors. These experiments were performed a minimum of five times. CD62P expression followed aggregation results. No statistically significant differences between platelets from normals and SCD patients with regard to ADP-induced CD62P expression were seen. All the donors lowered ADP-induced CD62P expression, and DEANO was again the only NO donor to significantly reduce ADP-induced CD62P expression in SCD platelets compared with normal platelets. OXINO slightly lowered the CD62P expression in both sets of platelets equally. CysNO yielded the largest amount of inhibition in both populations, but there was no statistically significant difference between the two groups. Again there was a large amount of variability in the responses of platelets to CysNO. When normal platelets were treated with ADP and CysNO, the percentage of CD62P-positive cells ranged from 11–80%; in platelets from SCD patients, these values ranged from 6–55%.
PAC-1 is an antibody specific for the active conformation of the platelet fibrinogen receptor, GPIIb–IIIa. We measured the binding of PAC-1 to platelets using flow cytometry (Fig 4). All the NO donors tested reduced the expression of the PAC-1 epitope by approximately 25–30%; however, we could not detect any statistically significant differences between platelets from normals and SCD patients.
We measured calcium elevation using the fluorescent dye fluo-4. When platelets were stimulated with thrombin, there was a large, immediate increase in fluorescence, with values decreasing and levelling off. The only difference between normal and SCD platelets was in the peak of fluorescence, with the SCD platelets having an increased response to thrombin. In four experiments, the mean fluorescence increased from 17 ± 9 to 120 ± 59 in normal platelets upon thrombin stimulation and from 30 ± 13 to 172 ± 65 in SCD patient platelets. The peak values measured immediately after thrombin addition differed in a statistically significant manner. However, when we measured calcium elevation induced by ADP with or without NO donors, there was no significant difference between normal and SCD platelets. In four experiments, the mean fluorescence increased from 17 ± 9 to 50 ± 25 in normal platelets and from 30 ± 13 to 66 ± 29 in SCD patient platelets. As platelets from SCD patients have been shown to be activated in the absence of agonist, aggregation assays were performed before calcium elevation measurements to determine the concentration of ADP at which SCD platelets exhibited increased aggregation over normal platelets. These concentrations ranged from 0·5 to 2 μmol/l and the resulting aggregation was decreased by approximately 60% (data not shown). This was done to maximize the differences in the activation state of the two platelet populations.
Erythrocytes are not the only blood cells affected by SCD; platelet activation occurs in SCD patients and may increase further during acute pain crisis (Browne et al, 1996; Wun et al, 1998). Activated platelets are thought to contribute to this process by secreting thrombospondin, CD62P and fibronectin from their α granules, thus providing the ligands for adhesion (Brittain et al, 1993). Studies have also measured increased plasma β-thromboglobulin (Mehta & Mehta, 1980), decreased platelet ADP/ATP ratios (Beurling-Harbury & Schade, 1989), increased urinary thromboxane metabolites (Kuranstin-Mills et al, 1994) and increased platelet procoagulant activity in SCD patients (Wun et al, 1998). Not only are platelet microparticles increased, but red cell vesiculation also adds to the presence of total microparticles (Allan et al, 1981). Thrombosis can be initiated by the stimulation of prothrombinase activity arizing from the increased exposure of negatively charged phospholipids (Kuypers et al, 1994).
Recently, the inhalation of NO has been considered as a potential therapy for SCD. NO is believed to improve red blood cell oxygen transport in SCD patients (Head et al, 1997; Kon et al, 1997). The effect of inhaled NO on platelets has also been studied and this treatment was found to be effective in reducing platelet aggregation in patients with acute respiratory distress syndrome (ARDS) (Samama et al, 1995). Our studies have focused on comparing the effects of NO on platelets from normal and SCD individuals. We were interested in determining what effects this potential therapy would have upon platelets, as they are key participants in thrombotic episodes. Three NO donors were selected to deliver NO to platelets. The first, DEANO, has been shown to produce pulmonary and systemic vasodilatation upon intravenous infusion (Vanderford et al, 1994). DEANO is a NO/nucleophile adduct that, when added to a solution of physiological pH, spontaneously releases two molecules of NO owing to the presence of an anionic nitrous dioxide (N2O2) group. DEANO has a half-life of approximately 2 min at 37°C and pH of 7·4 (Morley et al, 1993). The second donor, OXINO, is also a NO/nucleophile adduct and induces relaxation of vascular tissue in vitro and lowers blood pressure in vivo (Fukuto et al, 1992; Pino & Feelisch, 1994). OXINO is estimated to have a half-life lasting only seconds (Feelisch & Stamler, 1996). The third donor is CysNO, which is a nitrosothiol that participates in the metabolism of NO by transporting and targeting the NO group to specific thiol regulatory sites (Stamler, 1995). This compound has relaxant properties as well as antioxidant and antibacterial activities (Ignarro et al, 1981; Feelisch & Stamler, 1996). The half-life of CysNO in the absence of chelators, as in our experiments, is approximately 20–30 s (Feelisch & Stamler, 1996).
DEANO was the only NO donor that had significant differential effects on SCD and normal platelets. It significantly reduced ADP-induced platelet aggregation in platelets from SCD patients by approximately 45%. Aggregation of normal platelets was reduced by 25%. OXINO and CysNO both reduced the extent of ADP-induced aggregation equally in normal and SCD platelets. The same trend was seen in the experiments to measure CD62P expression. DEANO significantly reduced ADP-induced CD62P expression in SCD platelets from 60% to approximately 35%. Normal platelets showed only a 15% reduction. As in aggregation studies, OXINO and CysNO were equally effective in reducing CD62P expression in normal and SCD platelets. We also performed these studies with the NO donors present at a final concentration of 100 μmol/l (data not shown). There was no statistically significant difference between the resulting inhibition.
Platelet aggregation is the result of fibrinogen binding to its platelet receptor, GPIIb–IIIa. The fibrinogen binding site on this receptor is cryptic until a conformational change occurs and the binding sites are exposed (Phillips et al, 1988). This change is typically initiated by the signal transduction that accompanies platelet activation. PAC-1 is a monoclonal antibody that is used to measure the expression of this binding site, as it preferentially binds to the active conformation of GPIIb–IIIa (Shattil et al, 1985). We performed binding studies with this antibody to further explore the inhibition of platelet aggregation. Surprisingly, we did not detect differences between PAC-1 binding in normal and SCD platelets in the presence of DEANO, even though there were measurable differences in platelet aggregation. A previous study examined the effects of the NO donor SIN-1 (3-morpholino-synonimine) on normal platelets stimulated with thrombin (Keh et al, 1996). These investigators found that PAC-1 binding decreased upon the addition of 100 μmol/l SIN-1, but increasing concentrations of thrombin overcame the inhibition. When varying concentrations of SIN-1 were applied to thrombin-stimulated platelets, maximal inhibition of PAC-1 binding occurred between 10 and 100 μmol/l. Our studies were performed using ADP-stimulated platelets; this agonist was chosen as it has historically been used in platelet studies of sickle cell patients (Mehta & Mehta, 1980; Beurling-Harbury & Schade, 1989). Regardless of the agonist, we did see a decrease in PAC-1 binding upon addition of all three NO donors. We measured PAC-1 binding after 20 min; however, any differences in GPIIb–IIIa conformation between sickle and normal platelets may have a temporal basis such that 20 min was not sufficient for these differences to be measured. Our results suggest that the ability of SCD platelet GPIIb–IIIa to undergo conformational changes is not different from normal platelets or that the epitope for PAC-1 is not differentially expressed on normal and sickle cell patient platelets. The differences seen in platelet aggregation between these two populations probably results from alterations in other aspects of the signalling pathways utilized during platelet activation and subsequent aggregation.
Cytosolic calcium elevation is a hallmark of platelet activation (Rink & Sage, 1990). We compared the calcium elevation in SCD platelets to that of normal platelets and found that the only difference occurred during thrombin stimulation. Immediately after the addition of thrombin, there is typically a very large increase in calcium elevation that levels off over time. In SCD platelets, this increase was significantly larger than that measured in normal platelets, but the extent of calcium elevation decreased over time to a level equivalent to normal platelets. There were no differences between the two platelet populations when these cells were stimulated with ADP, either in the presence or absence of NO donors. These data suggest that the increased sensitivity to DEANO of platelets from SCD patients stimulated with ADP does not involve a change in calcium elevation. ADP is considered a ‘weak’ platelet agonist, especially when compared with thrombin. It is possible that the concentrations of ADP used were not sufficient to increase the calcium elevation. However, as these agonist concentrations induced platelet aggregation, it may be that the low platelet count and absence of plasma required to perform the flow cytometric assay were not optimum conditions for observing an increase in calcium elevation.
In conclusion, platelets from SCD patients stimulated by ADP are more sensitive to NO as delivered by DEANO. This sensitivity may be a result of two molecules of NO delivered spontaneously by DEANO as opposed to one molecule of NO from the other NO donors tested, or it may be a result of the longer half-life of DEANO. However, it is not the result of a quickly occurring conformational change in GPIIb–IIIa or calcium elevation. Further studies will focus on the mechanism of this sensitivity. The method of delivery of NO to platelets has varying effects upon platelet physiological responses and the utilization of the more effective delivery systems may contribute to reducing platelet activation in SCD patients.