Neal S. Kleiman, Methodist DeBakey Heart Center, 6565 Fannin, F1090, Houston, TX 77030, USA. Tel.: +1 713 441 4952; fax: +1 713 793 1352; e-mail: firstname.lastname@example.org
Background: The mechanisms for the variability in antiplatelet effects of aspirin are unclear. Immature (reticulated) platelets may modulate the antiplatelet effects of aspirin through uninhibited cyclooxygenase (COX)-1 and COX-2. Objectives: To evaluate the role of reticulated platelets in the antiplatelet effects of aspirin. Methods: Sixty healthy volunteers had platelet studies performed before and 24 h after a single 325-mg dose of aspirin. Platelet studies included light transmission aggregometry; P-selectin and integrin αIIbβ3 expression, and serum thromboxane B2 (TxB2) levels. Reticulated platelets and platelet COX-2 expression were measured using flow cytometry. Results: Subjects were divided into tertiles based on the percentage of reticulated platelets in whole blood. Baseline platelet aggregation to 1 μg mL−1 collagen, and postaspirin aggregations to 5 μm and 20 μm ADP and collagen, were greater in the upper than in the lower tertile of reticulated platelets. Stimulated P-selectin and integrin αIIbβ3 expression were also higher in the upper tertile both before and after aspirin. Platelet COX-2 expression was detected in 12 ± 7% (n = 10) of platelets in the upper tertile, and in 7 ± 3% (n = 12) of platelets in the lower two tertiles (P = 0.03). Postaspirin serum TxB2 levels were higher in the upper (5.5 ± 4 ng mL−1) than in the lower tertile (3.2 ± 2.5 ng mL−1, P = 0.03), and decreased even further with ex vivo additional COX-1 and COX-2 inhibition. The incidence of aspirin resistance (≥ 70% platelet aggregation to 5 μm ADP) was significantly higher in the upper tertile (45%) than in the lower tertile (5%, P < 0.0001). Conclusions: Reticulated platelets are associated with diminished antiplatelet effects of aspirin and increased aspirin resistance, possibly because of increased reactivity, and uninhibited COX-1 and COX-2 activity.
Aspirin is widely used in the primary and secondary therapy of atherosclerotic vascular disease. Aspirin inhibits formation of the vasoconstrictor and proaggregant thromboxane A2 (TxA2) by acetylating the enzyme prostaglandin H2 synthase (cyclooxygenase isoform 1, COX-1) at residue S530 . Despite an established 25% reduction in atherosclerotic events with aspirin use , recent observations have highlighted wide biological variability in the interindividual response to aspirin's antiplatelet effects. Observations have suggested associations between diminished responses to aspirin and thromboembolic events in a variety of clinical settings [3–5]. Understanding the mechanisms responsible for this phenomenon may be useful in devising appropriate therapy for patients whose biological response to aspirin is impaired.
Immature platelets, unlike mature platelets, contain mRNA and are termed ‘reticulated’ platelets, because of the staining patterns produced by the mRNA distribution. Detection of these platelets in the circulating blood reflects platelet production from megakaryoctes in the bone marrow and hence the rate of platelet turnover . In parallel, increased levels of circulating reticulated platelets have been reported in patients presenting with syndromes characterized by pathologic arterial thrombosis, including cerebrovascular events  or acute coronary syndromes . Several reports also indicate that resistance to the antithrombotic effects of aspirin is more prevalent in patients with strokes and coronary artery disease [5,9]. A high platelet turnover rate could theoretically produce a population of platelets that could confer resistance to aspirin through several different mechanisms. First, newly formed platelets might replenish platelet COX-1, thus allowing restoration of thromboxane synthesis in platelets not previously exposed to aspirin. A second mechanism through which immature platelets might permit aspirin-insensitive TxA2 synthesis is through uninhibited COX-2 activity. COX-2 is expressed in juvenile platelets, and can be detected in platelets of patients with high rates of platelet turnover [10,11]. Unlike COX-1, COX-2 is not inhibited by aspirin at oral doses of 81–325 mg daily. Although COX-2 does not ordinarily play a major role in thromboxane synthesis, it has been shown to contribute to TxA2 generation, particularly in endothelial cells [12–15]. We hypothesized that elevated platelet turnover, as assessed by reticulated platelets, is associated with increased platelet activity and diminished response to the antithrombotic effects of aspirin. As a corollary, we also evaluated whether uninhibited COX-1 and COX-2 contribute to thromboxane production in aspirin-treated individuals.
The study was approved by the Institutional Review Board at Baylor College of Medicine, and all subjects provided informed consent. In total, 60 healthy subjects not taking any other medications were recruited. Subjects had not taken non-steroidal anti-inflammatory drugs or aspirin for the preceding 2 weeks. Baseline blood samples were collected from an antecubital vein through a 21-gauge needle. After the collection of a baseline sample, subjects ingested a single dose of 325 mg of enteric-coated aspirin under supervision. A second blood sample was collected after 24 h.
Complete blood count
The complete blood count was measured using an automated cell-counting machine, Micros 60 (Horiba AX, Irvine, CA, USA).
Light transmission aggregometry (LTA)
Turbidimetric platelet aggregation was used to measure agonist-induced platelet aggregation. We collected 20 mL of whole blood in 0.32% sodium citrate (final concentration). The tubes were centrifuged immediately at 150 × g for 6 min to prepare platelet-rich plasma (PRP). The remaining whole blood fraction was further centrifuged at 1000 g for 15 min to separate platelet-poor plasma. The platelet count in PRP was standardized between 200 and 250 × 103μL−1. Platelet aggregation was induced using either 5 μm or 20 μm ADP, 1.5 mm arachidonic acid (AA) or 1.0 μg mL−1 collagen (final concentrations). The maximum aggregation achieved during a 6-min period was used for analysis. ‘Aspirin resistance’ was defined as ≥ 70% platelet aggregation in response to 5 μm ADP measured by LTA, using previously established criteria [4,9].
P-selectin and PAC-1
Platelet activation was determined by platelet surface expression of P-selectin and PAC-1 binding with whole blood flow cytometry as previously described . Briefly, integrin αIIbβ3 activation was assessed using a fluorescein isothiocyanate (FITC)-conjugated PAC-1 antibody (Becton Dickinson, San Jose, CA, USA), and P-selectin expression was determined using an R-phycoerythrin-conjugated anti-CD62P antibody (BD Pharmingen, San Jose, CA, USA). Five microliters of citrated whole blood was diluted with 70 μL of Tyrodes/bovine serum albumin followed by 5 μL of 20 μm ADP (10 μm final concentration) and then by 20 μL of anti-CD62b or PAC-1 antibody directed against integrin αIIbβ3. After incubation for 20 min, the mixture was fixed with phosphate-buffered saline (PBS) containing 1% paraformaldehyde. Samples were analyzed with a Coulter Epics XL MCL flow cytometer (Beckman Coulter, Miami, FL, USA). Unstimulated samples served as negative controls. Both PAC-1 binding and P-selectin were expressed as log mean fluorescence intensity (MFI) and as percentage change in MFI from baseline to the post-aspirin sample.
Reticulated platelets were measured by a previously described flow cytometry assay with slight modifications . Five microliters of whole blood was added to a tube containing 1 mL of thiazole orange (ReticCount, Becton Dickinson) and another tube with 1 mL of Isoflo (Beckman Coulter) as control. After incubation for 30 min in the dark, the tubes were spun at 1200 × g for 2.5 min to form a cellular pellet. The supernatant was discarded, and the pellet was resuspended in 1 mL of Isoflo. Within 1 h, flow cytometry was performed, with 5000 stained platelets being counted in the platelet gate. The platelets demonstrating an increase in the MFI beyond a threshold margin set at baseline were counted as reticulated platelets and expressed as a percentage of total platelets counted.
Platelet COX-2 expression was measured in 22 subjects, using flow cytometry as previously described . Briefly, prostaglandin I2 was added to PRP and centrifuged at 1500 × g for 10 min to obtain a platelet pellet. The pellet was then resuspended with equivalent CGS (13 mm sodium citrate, 30 mm glucose, 120 mm sodium chloride) solution, and centrifuged again at 1500 × g for 15 min, resuspended in Tyrodes solution, and fixed with paraformaldehyde 1% for 10 min. The platelets were then permeabilized by addition of Triton X-100 (0.3%) for 10 min, pelleted again at 12 000 g for 30 s, and resuspended in equivalent 1% PBS solution. Five microliters of platelets were incubated with 1 μg of FITC-conjugated anti-COX-2 antibody (Cayman Chemical, Ann Arbor, MI, USA) and analyzed on the flow cytometer. The platelet population was identified on the basis of its forward and side scatter distribution, and 10 000 platelets were analyzed.
Aggregation studies on isolated reticulated platelets
Aggregation of isolated reticulated and non-reticulated platelets was performed on platelets from two subjects. Reticulated platelets were separated from PRP using flow cytometry and thiazole orange staining. Platelets thus separated were then divided into two fractions; aspirin was added to one of these fractions to a final concentration of 50 μm. After incubation for 5 min, aggregation was induced with ADP 5 μm, collagen 2 μg mL−1 or AA 1.5 mm (final concentrations). As the number of reticulated platelets obtained was very low (< 15% of the total platelet count), LTA was not a feasible option. We measured aggregatory responses using flow cytometry in one subject and a Coulter counter in the other subject (Beckman Coulter, Fullerton, CA, USA). With the use of a single platelet size gate, only unaggregated platelets were counted; larger platelet aggregates were thus not counted. The number of platelets in an unstimulated sample was also measured. Aggregation was calculated by the formula (unstimulated count – stimulated count) ×100/unstimulated count.
Thromboxane B2 levels
Serum thromboxane B2 (TxB2) levels were measured by a commercially available enzyme-linked immunosorbent assay kit (Assay Designs, Ann Arbor, MI, USA) in sera collected by incubating whole blood at 37 °C in a glass tube without anticoagulants for 1 h and then centrifuging at 1000 g for 15 min to obtain the sera. TxB2 levels were measured at baseline, 24 h after aspirin administration, and after ex vivo COX-1 inhibition with 0.3 μm SC-560 (Cayman Chemicals, Ann Arbor, MI, USA) and COX-2 inhibition by 10 μm NS-398 (Cayman Chemicals), a specific COX-2 inhibitor. The IC50 value for SC-560 with respect to COX-1 is 9 nm, whereas the corresponding IC50 value for COX-2 is 6.3 μm . The IC50 values with NS-398 for human recombinant COX-1 and COX-2 are 75 and 1.77 μm respectively , and these concentrations of both inhibitors have been used in prior studies . Ex vivo inhibition with specific COX-1 and COX-2 inhibitors was performed to assess whether COX-1 and COX-2 are completely inhibited, and whether any further suppression of TxB2 could be achieved.
As previously planned, the subjects were divided into tertiles based on the percentage of reticulated platelets. Sample size was calculated to assess a 10% difference in LTA. An SD of 10% was assumed, and calculations were based on a power of 80% and a significance level of 0.05. Normally distributed data are expressed as mean ± SD. One-way anova was used to analyze differences in means between tertiles; a Kruskal–Wallis anova was used if the data were not normally distributed. A two-sided unpaired t-test was used to compare means between two groups; the Mann–Whitney test was used if the data were not normally distributed. Fisher's exact test was used to compare the upper tertile with the other two tertiles, for dichotomous variables. Linear regression analysis was also performed, and correlation coefficients were calculated for relationships between reticulated platelet count and agonist-induced platelet aggregation. P < 0.05 was considered to be statistically significant.
Subjects in each of the three tertiles did not differ in age, sex or body mass index (Table 1). Other hematologic parameters were also similar in the three groups, except for mean platelet volume, which was significantly higher in the upper two tertiles, reflecting the increased size of reticulated platelets.
Table 1. Baseline characteristics in tertiles of % reticulated platelets
*anova between tertiles.
BMI, body mass index; WBC, white blood cell count; MPV, mean platelet volume; NS, not significant.
31 ± 5
34 ± 7
31 ± 3
BMI (kg m−2)
23 ± 2.3
25 ± 4
23 ± 6.3
Reticulated platelets (% of total platelets)
12 ± 2
8 ± 1
4 ± 1
WBC (103 mm−3)
5.1 ± 0.9
5.4 ± 1.2
4.9 ± 0.9
Hemoglobin (g dL−1)
11.4 ± 1.4
11.8 ± 1.6
11.5 ± 1.4
Platelets (103 mm−3)
215 ± 96
221 ± 64
231 ± 54
7.4 ± 0.7
7.5 ± 0.5
6.9 ± 0.6
Baseline platelet aggregation in response to 5 μm ADP was slightly higher in the upper tertile than in the lower tertile of reticulated platelets (Table 2), with borderline significance (P = 0.06). Baseline amounts of aggregation to 20 μm ADP and 1.5 mm AA were similar in the upper and lower tertiles. The response to 1 μg mL−1 of collagen before aspirin was greater in the upper tertile (80 ± 7%) than in the lower tertile (69 ± 22, P = 0.03). After aspirin was administered, platelet aggregation in response to both 5 and 20 μm ADP and collagen was greater among subjects in the upper tertile than among those in the lower tertile (Table 2, Fig. 1).
Table 2. Platelet aggregation measured by light transmittance aggregometry in tertiles of % reticulated platelets
Baseline platelet aggregation (%)
Postaspirin platelet aggregation (%)
*Upper vs. lower tertile, Student's t-test.
ADP 5 μm
83 ± 14
84 ± 12
72 ± 20
67 ± 9
60 ± 8
54 ± 11
ADP 20 μm
90 ± 4
91 ± 4
86 ± 10
81 ± 6
80 ± 7
75 ± 7
Arachidonic acid 1.5 mm
72 ± 28
64 ± 32
67 ± 29
8 ± 7
7 ± 5
5 ± 4
Collagen 1.0 μg mL−1
80 ± 7
77 ± 16
69 ± 22
24 ± 8
15 ± 8
8 ± 4
Correlation of percentage reticulated platelets and platelet aggregation
Percentage reticulated platelets significantly correlated with platelet aggregation in response to 5 μm ADP (r = 0.53, P < 0.0001) and 1 μg mL−1 collagen (r = 0.71, P < 0.0001; Fig. 2).
Surface markers of platelet activation
Baseline P-selectin expression and PAC-1 binding after stimulation with 10 μm ADP were significantly higher among subjects in the upper tertile than among those in the lower tertile (Table 3). These differences continued even after aspirin administration. However, aspirin failed to decrease expression of either P-selectin or integrin αIIbβ3 by a significant degree in any of the tertiles (Table 3).
Table 3. P-selectin and PAC-1 expression expresses as mean florescence intensity (MFI) in tertiles of % reticulated platelets
*Upper vs. lower tertile, Student's t-test.
17.8 ± 8.2
15.7 ± 5.7
12.6 ± 3.3
16.7 ± 7.2
15.9 ± 4
12.9 ± 4.3
9.4 ± 2.5
8.7 ± 2.4
7.5 ± 2.7
8.9 ± 2.3
8.9 ± 1.8
6.7 ± 2.3
Platelet COX-2 expression was measured randomly in 10 subjects from the upper tertile and 12 subjects from the remaining two tertiles. COX-2 was detected in 12 ± 7% of platelets in the upper tertile, and in 7 ± 3% of platelets in the lower two tertiles (P = 0.03).
Serum TxB2 levels prior to aspirin administration were not different between the tertiles. After aspirin administration, serum TxB2 levels decreased in all subjects, but remained significantly higher in the upper tertile than in the lower tertile (5.5 ± 4 ng mL−1 vs. 3.2 ± 2.5 ng mL−1, P = 0.03) (Table 4). We further tested, ex vivo, the contribution of uninhibited COX-1 (using 0.3 μm SC-560) and COX-2 (using 10 μm NS-398) activity to TxB2 production in postaspirin blood samples. Selective inhibition of COX-1 with SC-560 reduced TxB2 levels significantly (Table 4). Similarly, selective inhibition of COX-2 with NS-398 reduced TxB2 levels, but to a slightly lesser degree than SC-560 (82% vs. 90% respectively, P = 0.05). Importantly, addition of either inhibitor ex vivo eliminated the observed differences in TxB2 levels between the tertiles.
Table 4. Serum thromboxane B2 (TxB2, ng mL−1) synthesis in tertiles of % reticulated platelets
*Upper vs. lower tertile, Mann–Whitney test.
417 ± 157
387 ± 206
351 ± 195
5.5 ± 4
4.8 ± 2.7
3.2 ± 2.5
Postaspirin TxB2 with ex vivo SC-560
0.49 ± 0.46
0.49 ± 0.52
0.35 ± 0.30
Postaspirin TxB2 with ex vivo NS-398
1.18 ± 1.46
0.43 ± 0.27
0.65 ± 0.69
Platelet aggregation studies with isolated reticulated platelets
In vitro aspirin decreased aggregation in response to ADP by 17% in isolated non-reticulated platelets and by 4% in isolated reticulated platelets. The response to collagen decreased by 19% in non-reticulated platelets and by 12% in reticulated platelets. The response to AA decreased by 23% in non-reticulated platelets and by 7% in reticulated platelets (Fig. 3).
Aspirin resistance (defined by aggregation in response to 5 μm ADP ≥ 70%) was present in 45% (9/20) in the upper tertile, 10% (2/20) in the middle tertile, and 5% (1/20) in the lower tertile (P < 0.0001, Fig. 4).
This study demonstrates for the first time that high levels of circulating immature platelets, possibly reflecting a high rate of platelet turnover, may mediate resistance to the antithrombotic effects of aspirin. The findings are internally concordant, in that we observed, in subjects with high levels of reticulated platelets, diminished responses to the effects of aspirin detected by LTA, increased expression of surface markers of platelet activation, and increased TxB2 production. Furthermore, our results suggest that when reticulated platelets are isolated, the effects of aspirin in these platelets are less than those observed in mature platelets.
It is possible that high platelet turnover confers diminished aspirin response because of increased generation of uninhibited COX-1. We tested this concept by measuring serum TxB2 after additional COX-1 inhibition of postaspirin blood samples with a specific COX-1 inhibitor, SC-560. Incremental inhibition of COX-1 decreased serum TxB2 levels significantly, and obviated the differences in TxB2 synthesis between tertiles of platelet turnover, suggesting incomplete inhibition of COX-1 with aspirin, and more so in conditions of high platelet turnover.
We further tested whether low response to aspirin may be explained by non-COX-1-mediated thromboxane production. Although platelets contain COX-1 predominantly, about 8–10% of platelets express COX-2 [10,11]. In patients with high platelet turnover states (such as idiopathic thrombocytopenic purpura [ITP], multiple myeloma, and following cardiac surgery), COX-2 is expressed in 30–60% of platelets [11,20]. Zimmermann et al. studied patients immediately following cardiac surgery; by day 5, COX-2 expression increased sixteenfold from preoperative levels, presumably because of increased platelet release from the bone marrow. In concordance with this study, we demonstrated that platelet COX-2 expression is greater in subjects with higher levels of reticulated platelets than in patients with lower levels of reticulated platelets.
It was not clear whether the increased platelet COX-2 expression plays a role in TxB2 synthesis or platelet aggregation. In platelets with COX-2 expression, in vitro inhibition with NS-398 decreased TxB2 synthesis significantly . In contrast with these results, in patients who underwent cardiac bypass surgery, COX-2 inhibition with celecoxib did not decrease thromboxane synthesis . More recently, in smokers, extraplatelet COX-2 but not platelet COX-2 was suggested as a contributor to thromboxane synthesis, in the absence of aspirin . Hence, to better understand the role of COX-2 activity in TxB2 synthesis in the presence of aspirin, we measured serum TxB2 levels after ex vivo inhibition of COX-2 in postaspirin blood samples. TxB2 synthesis was significantly suppressed by ex vivo COX-2 inhibition, suggesting that uninhibited COX-2 activity may contribute to aspirin-independent TxB2 synthesis, particularly in conditions of high platelet turnover.
The clinical impact of reticulated platelets and platelet turnover on responses to aspirin will need to be studied in a large clinical setting. In patients hospitalized for cerebrovascular events, the percentage of reticulated platelets was significantly greater than in controls subjects without strokes . Among patients with acute coronary events, those with an acute myocardial infarction had a significantly greater reticulated platelet count when compared with patients with unstable angina or stable angina . In these studies, the increased platelet turnover could be a consequence of the acute event. Whether an increased platelet turnover is a causative factor in acute ischemic cardiac or cerebrovascular events is not known. Our study has demonstrated that a higher reticulated platelet count in healthy subjects diminishes platelet responses to aspirin, and could thus be relevant in a clinical setting as well.
Although there is serious concern regarding COX-2 inhibition in patients with coronary artery disease [21,22], the role of COX-2 antagonism, possibly transiently, in suppressing thromboxane synthesis should be studied further. We did not characterize the origin of TxB2 production, which could have both platelet and extraplatelet sources. Monocytes and endothelial cells express COX-2 and demonstrate the ability to synthesize thromboxane [12–14,23]. Thus, the contribution of other cell types still needs to be assessed.
In conclusion, aspirin resistance is associated with increased levels of reticulated platelets. Increased reactivity of younger platelets, increased COX-2 expression and activity and incompletely inhibited COX-1 may be possible mechanisms for aspirin resistance. Mechanisms by which reticulated platelets and platelet turnover impact on coronary events, and methods to modify antiplatelet therapy on the basis of platelet turnover, will need to be further evaluated.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.