Platelet hyperreactivity generalizes to multiple forms of stimulation

Authors

  • D. L. YEE,

    1. Department of Pediatrics, Hematology-Oncology Section
    2. Department of Medicine, Thrombosis Research Section, Baylor College of Medicine, One Baylor Plaza, Houston, TX, USA
    Search for more papers by this author
  • A. L. BERGERON,

    1. Department of Medicine, Thrombosis Research Section, Baylor College of Medicine, One Baylor Plaza, Houston, TX, USA
    Search for more papers by this author
  • C. W. SUN,

    1. Department of Medicine, Thrombosis Research Section, Baylor College of Medicine, One Baylor Plaza, Houston, TX, USA
    Search for more papers by this author
  • J.-F. DONG,

    1. Department of Medicine, Thrombosis Research Section, Baylor College of Medicine, One Baylor Plaza, Houston, TX, USA
    Search for more papers by this author
  • P. F. BRAY

    1. Department of Pediatrics, Hematology-Oncology Section
    2. Department of Medicine, Thrombosis Research Section, Baylor College of Medicine, One Baylor Plaza, Houston, TX, USA
    Search for more papers by this author

Paul F. Bray, Thrombosis Research Section, Baylor College of Medicine, One Baylor Plaza, BCM 286, N1319, Houston, TX 77030, USA.
Tel.: +1 713 798 3480; fax: +1 713 798 3415; e-mail: pbray@bcm.tmc.edu

Abstract

Summary. Background: Although platelet hyperreactivity constitutes an important cardiovascular risk factor, standardized methods for its measurement are lacking. We recently reported that aggregometry using a submaximal concentration of epinephrine identifies individuals with in vitro platelet hyperreactivity; this hyperreactivity was reproducible on multiple occasions over long periods of time. Objective and methods: To better understand this aberrant reactivity, we studied in a large group of subjects (n = 386) the relationship between healthy individuals’ platelet reactivity to epinephrine and their platelet phenotype as measured by other functional assays. Results: Subjects with hyperreactivity to epinephrine were more likely to exhibit hyperfunction in each major aspect of platelet activity, including adhesion (response to low-dose ristocetin; P < 0.001), activation (surface P-selectin expression and PAC-1 binding after stimulation; P ≤ 0.003) and aggregation to other agonists [no agonist, adenosine diphosphate (ADP), arachidonic acid, collagen, collagen-related peptide and ristocetin; P ≤ 0.025] and to applied shear stress (PFA-100 and cone-and-plate viscometer; P < 0.05). These differences persisted after adjusting for demographic and hematologic differences between groups. We studied candidate genes relevant to epinephrine-mediated platelet activation and found that hyperreactivity to epinephrine was associated with a polymorphism on the gene (GNB3) encoding the beta-3 subunit of G proteins (P = 0.03). Conclusions: Robust aggregation to a submaximal concentration of epinephrine establishes a true hyperreactive platelet phenotype that is ‘global’ as opposed to agonist specific; detection of this phenotype could be useful for studying patients at risk for arterial thrombosis. The mechanisms underlying hyperreactivity to different types of platelet stimulation may share common signaling pathways, some of which may involve specific G protein subunits.

Introduction

Platelets are integral components of occlusive thrombi that result in myocardial infarction and ischemic stroke; they may also contribute to the development and progression of more chronic atherosclerotic lesions that underlie these acute events [1]. It is therefore not surprising that numerous prospective studies link increased platelet function (i.e. platelet hyperreactivity) with worse clinical outcomes in various cardiovascular settings [2–7]. However, widespread recognition of platelet hyperreactivity as an important clinical risk factor has been limited by the lack of a standardized definition as well as by technical challenges posed by platelet function testing for the typical clinical laboratory. These limitations have also hampered research progress in this area, as the variety of platelet assays used by different investigators has led to difficulties in study comparison and interpretation.

Turbidometric platelet aggregometry using platelet-rich plasma (PRP) is widely used for the assessment of platelet function and remains the ‘gold standard’ according to many experts [8,9] despite its reported limitations. We recently showed that aggregometry using a submaximal concentration of epinephrine as agonist (0.4 μm in citrated PRP) reliably identifies a distinct subgroup of healthy individuals with in vitro platelet hyperreactivity [10]. This hyperreactivity was reproducible over long periods of time (several years), suggesting that platelet phenotype as determined by this assay is intrinsic for a given individual. However, the physiologic basis for and significance of this finding remain unclear. In the present study, to better understand aberrant platelet reactivity to epinephrine, we characterized this hyperreactive platelet phenotype using multiple other functional and biochemical assays that measure a broad spectrum of physiologic responses. In an effort to explore potential mechanisms responsible for platelet hyperreactivity to epinephrine, we also measured expression of key surface membrane molecules and genotyped polymorphisms located on candidate genes relevant to epinephrine-mediated platelet activation. Importantly, our study of a large number of healthy individuals (n = 386) enabled us to consider the effects of numerous variables on platelet reactivity and to more accurately assess their impact by adjusting for key differences between subjects.

Methods

Subject recruitment and sample collection

This study was approved by the Baylor College of Medicine institutional review board. Between July 2001 and July 2005, healthy subjects (an expanded version of our previously described cohort [10]) were recruited through printed advertisements. Subjects were excluded if they were taking medications known to affect platelet function (including non-steroidal anti-inflammatory drugs during the 48 h prior to phlebotomy and aspirin) or more than one prescription medication. After informed consent was obtained, subjects were asked to fast overnight and to refrain from intensive exercise and tobacco use for 4 h prior to an early morning phlebotomy. After resting comfortably for at least 10 min, subjects were phlebotomized. After the first 2 mL were discarded, blood was collected through a 19-gauge needle into a syringe containing 3.8% sodium citrate in a prespecified volume to ensure a 9:1 whole blood/citrate ratio. Some subjects had additional blood collected into syringes containing the thrombin inhibitor Phe-Pro-Arg-chloromethyl ketone (PPACK, 75 μm final concentration). Demographic information was recorded for each subject based on information obtained via direct interview and a questionnaire.

Reagents

All reagents were obtained and careful quality control measures were taken as previously described [10].

Platelet aggregometry

Plasma (undiluted) platelet count and mean platelet volume (MPV) were measured using a Z2 Coulter Counter (Beckman-Coulter, Fullerton, CA, USA). As previously described in detail [10], undiluted plasma was diluted with autologous platelet-poor plasma to generate standardized PRP (between 200 000 and 250 000 platelets μL−1) on which platelet aggregometry was performed using a Bio/Data 4-channel platelet aggregometer (Horsham, PA, USA) and the following agonists at a range of concentrations: epinephrine, ADP, collagen, ristocetin and collagen-related peptide. The primary variable of interest was maximal aggregation (%). PRP from each subject was assayed using 0.5 mg mL−1 arachidonic acid; 99% of subjects exhibited aggregation to this agonist, confirming the low frequency of aspirin use in our cohort. All aggregation studies were completed within 2 h of phlebotomy.

Studies of platelet surface molecule expression

At room temperature, whole blood aliquots were treated with the peptide Gly-Pro-Arg-Pro amide acetate in a 9:1 ratio then diluted (1:17) with calcium- and magnesium-free Tyrode's buffer containing 1% bovine serum albumin (BSA) prior to incubation with ADP (final concentrations 0–20 μm) for 5 min. Antibody [either phycoerythrin-conjugated monoclonal anti-CD62P (AK4) or fluorescein isothiocyanate-conjugated PAC-1 (BD-PharMingen, San Diego, CA, USA)] at a final concentration of 5 μg mL−1 was added followed by incubation for 20 min. Tubes containing anti-CD62P were fixed with phosphate-buffered saline containing 1% paraformaldehyde (PBS/1% PFA) and tubes containing PAC-1 antibody were treated with Tyrode's buffer/1% BSA. After 2 min, the samples were analyzed on a Coulter Epic XL MCL flow cytometer (Beckman-Coulter) by first identifying the platelet population based on the particle size on forward scatter and then collecting data on 10 000 events from the defined platelet gate. PAC-1 binding and P-selectin expression were quantified as units of geometrical mean fluorescence. Additional whole blood specimens were treated with peptide as above before incubation with one of four phycoerythrin-conjugated monoclonal antibodies used to measure expression of the surface receptors FcγRIIA, GPIIb-IIIa, GPIb-IX-V and GPIa-IIa: 2E1 (anti-CD32; Beckman-Coulter), HIP8 (anti-CD41a; BD-Pharmingen), SZ2 (anti-CD42b; Beckman-Coulter) and AK7 (anti-CD49b; Research Diagnostics, Concord, MA, USA), respectively. Incubation in Tyrode's buffer/1% BSA proceeded for 20 min before fixation with PBS/1% PFA. Mean fluorescence intensity was measured as above. Specimens treated with phycoerythrin-conjugated goat antimouse IgG served as controls.

Shear-induced platelet activation and aggregation

As previously described [11], PRP was exposed to various levels of shear stress for 60 s on a computerized cone-and-plate viscometer (RS1; HAAKE Instrument Inc., Paramus, NJ, USA). Aliquots were then diluted (1:9) with Tyrode's buffer/1% BSA, stained for surface P-selectin and incubated; PBS/1% PFA was added and the samples were analyzed for P-selectin expression as above. Unsheared PRP was the negative control. For the PFA-100 studies, whole blood specimens were aliquoted into the analyzer reservoir in accordance with the manufacturer's instructions (Dade-Behring, Deerfield, IL, USA) and the closure time (seconds) was measured. Testing was performed using cartridges coated with collagen/epinephrine and collagen/ADP.

Other plasma assays

Fibrinogen concentration using the Clauss method and von Willebrand factor (VWF) activity using a ristocetin cofactor assay were measured on a Dade-Behring BCS machine according to the manufacturer's instructions.

Gene association studies

Candidate gene polymorphisms were genotyped using the real-time polymerase chain reaction (PCR) on a 7700 Sequence Detector (Applied Biosystems, Foster City, CA, USA). Genomic DNA was extracted from buffy coats using a Qiagen DNA extraction kit (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions. For each PCR reaction, 100 ng of genomic DNA was mixed with 2.5 pm of two fluorescent probes, and two primers spanning the polymorphic site. The fluorescent probes were allele specific and tagged with two different reporter dyes (VIC and FAM). Because various numbers of tandem repeats of 39 bases constitute the VNTR polymorphism of GPIbα, specimens were genotyped by size exclusion of PCR amplified DNA fragments using a primer pair. Primers and probes used for genotyping are listed in Table 1. The basic PCR reaction was set at 40 cycles with denaturing for 15 s at 95 °C and annealing/extension for 60 s at 65 °C, with minor optimizing modifications for each polymorphism. To ensure accuracy, 5% of samples were randomly chosen for repeat genotyping using restriction digestion of polymorphism-containing DNA fragments generated through PCR DNA amplification; the match rate was 100% for each polymorphism.

Table 1.   Primer and probe sequences for genotyping
PolymorphismMoleculeFluorescent probe
G1838AA2ARCCCAACTCTCTCTCTCTTTTTG/AAAGAAAAATGC
C825T (GNB3)β3 subunitCATCACGTCC/TGTGGCCTTCTCC
L33P (PLA)GPIIIaCCT GCC TCT/CGGG CTC ACC TC
T145M (Ko)GPIbαAGGGCTCCTGAC/TGCCCACACC
VNTRGPIbαForward primer: AGGACTGTGGTCAAGTTCCC
Reverse primer: CGGAGCTTTGGTGGCTGATC
C-5T (Kozak)GPIbαTGCCCACAGGC/TCCTCATGC
C807TGPIaACCTCACAAACACATTC/TGGAGCAATTCA
E505K (Br)GPIaTACTATCAAAG/AAGGTAAAAA
T13254CGPVIACCACCTTCCC/TCGGTA
L843S (Bak)GPIIbAGGGGCTGGGGA/CTGGGCAG
H131RFcγRIIACAGAAATTCTCCCA/GTTTGGATCCCACCT

Statistical analysis

SPSS version 12.0 (SPSS Inc., Chicago, IL, USA) was used for all data analyses. All data were examined using histograms and scatter plots. The chi-squared test and analysis of variance were used to compare categorical and continuous variables, respectively, between groups. The Mann–Whitney U-test was used when the distribution was not normal. Linear and logistic regression models were employed to adjust for differences in covariates between groups. Correlation between two continuous variables was calculated using Pearson's correlation coefficient. The chi-squared test was also used to compare allele frequencies with the Hardy–Weinberg equilibrium prediction. P-values < 0.05 were considered significant.

Results

Platelet hyperreactivity to epinephrine

We established platelet phenotypes by studying platelet function in healthy adults under uniform conditions. Figure 1 depicts the aggregation response to 0.4 μm epinephrine in PRP of our study cohort of 386 healthy subjects. Although the vast majority of individuals exhibit relatively little platelet aggregation (< 40%), a substantial minority (14%) demonstrates > 60% aggregation, consistent with a hyperreactive phenotype. For the purposes of this study, we thus classified individuals with at least 60% aggregation to 0.4 μm epinephrine as hyperreactive to epinephrine.

Figure 1.

 Distribution of aggregation response to submaximal epinephrine concentration among healthy individuals. This histogram depicts the number of subjects (y-axis) with a given level of platelet aggregation (x-axis) to 0.4 μm epinephrine. Fifty-five out of 386 subjects (14%) showed an unusually robust aggregation response (≥ 60%) and were the focus of this study.

Comparison of subjects’ demographic and hematologic characteristics

Based on the distinct bimodal distribution that characterized the range of responses to this low concentration of epinephrine (Fig. 1), we compared demographic and hematologic parameters between subjects whose platelets demonstrated at least 60% aggregation on this assay and those who did not (Table 2). The two groups were similar according to our collected demographic data, except that the hyperreactive group had a higher percentage of women (P < 0.02). Among these female subjects, we found no association between the hyperreactive response and oral contraceptive use, menopausal status, or phase of the menstrual cycle at the time of testing. With respect to hematologic parameters, subjects who were hyperreactive to epinephrine had lower undiluted PRP platelet counts (P = 0.004), higher MPVs (P < 0.001) and higher plasma fibrinogen levels (P = 0.009). There was no difference in hematocrit or VWF activity between the two groups (Table 2).

Table 2.   Subject demographic and hematologic characteristics by response to 0.4 μm epinephrine
Subject characteristic< 60% aggregation (n = 331)At least 60% aggregation (n = 55)P
  1. Data expressed as mean ± SD or percentages. Only one subject (who had aggregation < 60%) had diabetes. PRP, platelet-rich plasma.

Age (years)33.4 ± 10.133.1 ± 8.80.836
Women (%)61.378.20.016
Race (% of group)
 African–American35.047.30.370
 White37.527.3
 Hispanic15.710.9
 Asian11.214.5
 Other0.60.0
Body mass index, kg m−226.4 ± 5.626.7 ± 6.10.645
Women on oral contraceptives (%)25.620.90.518
Women, premenopausal (%)84.686.00.807
Premenopausal women in luteal phase (%)56.171.00.129
Current smokers (%)11.97.30.319
With hypertension (%)6.100.06
Fibrinogen (mg dL−1)319 ± 88354 ± 1020.009
Hematocrit (%)37.2 ± 4.436.7 ± 5.00.514
Mean platelet volume (MPV) (fL)7.00 ± 0.647.43 ± 0.77< 0.001
Undiluted PRP platelet count (103 μL−1)427 ± 102384 ± 860.004
Ristocetin cofactor activity (%)83.1 ± 3689.4 ± 360.252

Platelet response to alternate agonists

The distribution pattern of aggregation response to submaximal concentrations of ADP, collagen, collagen-related peptide and ristocetin was bimodal [10], similar to that shown in Fig. 1, with a distinct minority of healthy individuals exhibiting > 60% aggregation and the remainder exhibiting markedly less. In order to assess the relationship between platelet hyperreactivity to epinephrine and hyperreactivity to these other agonists, as we did for epinephrine (and based on the distinctly bimodal distributions observed [10]), we classified individuals with at least 60% aggregation to a particular agonist as hyperreactive to that agonist. Table 3 shows that hyperreactivity to epinephrine was strongly associated with hyperreactivity to submaximal concentrations of each of the other agonists tested (P ≤ 0.025), with odds ratios (estimating probabilities of hyperreactivity to each agonist based on reactivity to submaximal epinephrine) ranging from 3.1 (for collagen-related peptide) and 36.1 (for ADP). Except for collagen-related peptide, these associations persisted even after adjusting for differences in gender, fibrinogen, PRP platelet count and MPV between the two groups (P ≤ 0.001, estimated odds ratios between 3.6 and 23.0). Hyperreactivity to epinephrine was associated with an unusually robust response to even a very low concentration of ristocetin (0.5 mg mL−1), a condition commonly used to detect increased binding between platelets and VWF [12], a critical step in platelet adhesion. From a smaller group of subjects (n = 228) who had separate additional specimens anticoagulated with PPACK, we observed a strong association (adjusted OR = 21.5) between hyperreactivity to epinephrine in citrated vs. PPACK-anticoagulated specimens.

Table 3.   Association between platelet hyperreactivity to epinephrine and hyperreactivity to other agonists
Agonist, concentrationOR*PORP
  1. All studies performed using citrated platelet-rich plasma except for 10 μm epinephrine (PPACK). *OR = odds ratio = probability of hyperreactivity to the specified agonist given hyperreactivity to epinephrine/probability of hyperreactivity to the specified agonist given no hyperreactivity to epinephrine, along with 95% confidence interval. Adjusted for gender, fibrinogen, mean platelet volume, platelet count. Adjusted OR not calculated as a result of the small number of subjects.

ADP, 1 μm36.1, [13.7, 95.4]< 0.00123.0, [8.1, 65.1]< 0.001
Collagen, 20 μg mL−14.3, [2.2, 8.5]< 0.0013.6, [1.7, 7.7]0.001
Collagen-related peptide, 0.005 μg mL−13.1, [1.1, 8.7]0.0252.5, [0.6, 9.9]0.188
Ristocetin, 0.5 mg mL−17.9, [6.0, 10.4]< 0.001
Ristocetin, 0.75 mg mL−14.1, [2.2, 7.7]< 0.0014.2, [2.1, 8.3]< 0.001
Epinephrine, 10 μm29.5, [5.9, 146.7]< 0.00121.5, [4.0, 116.1]< 0.001

In contrast to the above agonists, addition of no agonist (spontaneous aggregation) yielded a fairly normal, unimodal distribution of responses in our cohort (data not shown). We compared the extent of spontaneous aggregation between subjects based on their response to 0.4 μm epinephrine (Fig. 2) and observed that spontaneous aggregation was higher in the hyperreactive group (P = 0.001, adjusted P = 0.012). Aggregation to 0.5 mg mL−1 arachidonic acid was also higher in the hyperreactive group (P < 0.001, adjusted P < 0.001).

Figure 2.

 Comparison of aggregation response. Using platelet-rich plasma (PRP) aggregometry, response to no agonist (A) and to 0.5 mg mL−1 arachidonic acid (B) was higher in the group with hyperreactivity to 0.4 μm epinephrine (≥ 60% aggregation, n = 50, black bars) compared to all others (< 60% aggregation, n = 303, gray bars).

We measured P-selectin expression on the platelet surface (a marker of platelet activation and alpha-granule secretion) [13] after stimulation with varying concentrations of ADP (Fig. 3). At each concentration tested, surface P-selectin expression was higher in the subjects with hyperreactivity to epinephrine (P ≤ 0.001, adjusted P ≤ 0.002).

Figure 3.

 Comparison of platelet surface P-selectin expression after adenosine diphosphate (ADP) simulation. Data are expressed as difference in mean fluorescence intensity between stimulated and unstimulated condition, except at concentration = 0, where unsubtracted data are presented. Data available for 347 subjects categorized by response to 0.4 μm epinephrine [49 with ≥ 60% aggregation (black bars), 298 with < 60% (gray bars)]. Values above bar pairs represent P-values.

Platelet response to shear stress

Because shear stress is an important stimulus for platelet thrombus formation [14], we studied platelet response to shear using two different assay systems. Subjects with platelet hyperreactivity to epinephrine had shorter PFA-100 closure times than those who did not (Fig. 4A). This occurred with both ADP (P < 0.05, adjusted P = 0.03) and epinephrine (P < 0.001, adjusted P < 0.001) cartridges. We conducted further testing under shear conditions by examining P-selectin expression after applying shear stress to subjects’ PRP specimens using a cone-and-plate viscometer. After exposure to shear rates of 1000, 2000 and 10 000 s−1, platelets from subjects with hyperreactivity to epinephrine showed greater P-selectin expression (Fig. 4B). These differences were not observed after adjusting for the differences in fibrinogen level, PRP platelet count and MPV between groups (adjusted P > 0.4).

Figure 4.

 Comparison of platelet response to shear stress. Data are categorized by subjects’ responses to 0.4 μm epinephrine [gray bars represent subjects with < 60% aggregation, black bars represent subjects with ≥ 60% aggregation (epinephrine hyperreactivity)]. (A) Mean PFA-100 closure times in subjects’ whole blood specimens tested with the adenosine diphosphate (ADP) cartridge (data available for 346 subjects, 51 with epinephrine hyperreactivity) and with the epinephrine cartridge (data available for 312 subjects, 45 with epinephrine hyperreactivity). (B) Mean platelet surface P-selectin expression at different shear rates. Data are expressed as difference in mean fluorescence intensity between stimulated and unstimulated condition. Values above bars represent P-values. Data available for 345 subjects (49 with epinephrine hyperreactivity).

Expression of platelet surface receptor molecules

Platelet membrane glycoproteins (GP) are central effectors of platelet adhesion, activation and aggregation. Accordingly, we assessed whether differences in expression of these platelet surface receptors were associated with the hyperreactive platelet phenotype. Surface expression of the platelet fibrinogen receptor (GPIIb-IIIa) was measured under resting and stimulated conditions. Both GPIIb-IIIa expression (on unstimulated platelets) and PAC-1 antibody binding (a measure of expression of the activated form of GPIIb-IIIa) after stimulation with ADP were greater in subjects with hyperreactivity to epinephrine (Fig. 5). The differences in PAC-1 binding persisted even after adjusting for differences in gender, fibrinogen levels, PRP platelet count and MPV (adjusted P ranged from 0.01 to 0.04); however the differences in GPIIb-IIIa expression did not. FcγRIIA expression was higher in the hyperreactive group (P = 0.009), but this difference also resolved after adjusting for differences in MPV. We observed no difference between subjects in surface expression of the receptor molecules GPIb-IX-V or GPIa-IIa (data not shown).

Figure 5.

 Comparison of platelet fibrinogen receptor expression and activation. Data are available for 347 subjects and are categorized by subjects’ responses to 0.4 μm epinephrine [49 with ≥ 60% aggregation (hyperreactive-black bars), 298 with < 60% (gray bars)]. (A) GPIIb-IIIa expression as measured by flow cytometry. (B) PAC-1 binding after adenosine diphosphase (ADP) stimulation at varying concentrations. Data are expressed as difference in mean fluorescence intensity (MFI) between stimulated and unstimulated condition, except at concentration = 0, where unsubtracted data are presented. Values above bars represent P-values.

Gene association studies

Because the α2A-adrenergic receptor (A2AR) mediates platelet activation by epinephrine [15,16], we examined the potential role of two gene polymorphisms located on candidate genes relevant to this receptor. We found no association (P = 0.654; Table 4) between platelet hyperreactivity to epinephrine and the G1838A polymorphism (V. Afshar-kharghan and P. F. Bray, Baylor College of Medicine, Houston, TX, USA; unpublished data) on the A2AR gene which has been linked to variation in epinephrine-mediated platelet aggregation [17]. However, platelet hyperreactivity to epinephrine was associated with the C825T polymorphism located on the gene encoding the β3 subunit of G-proteins (GNB3; P = 0.03; Table 4). Importantly, we found no association between platelet hyperreactivity to epinephrine and any of nine other polymorphisms located on genes encoding platelet surface molecules physiologically distinct from the A2AR (Table 1). The allele frequencies of each polymorphism were in Hardy–Weinberg equilibrium, with the exception of the C825T and T13254C polymorphisms. The allele frequencies for these polymorphisms differed markedly between ethnic groups (data not shown); when analyzed by the ethnic groups described in Table 2, these data were consistent with the Hardy–Weinberg prediction.

Table 4.   Association between platelet hyperreactivity to epinephrine and genotypes for A2AR and GNB3
 < 60% aggregationAt least 60% aggregationP
  1. Data (available for 386 subjects) expressed as number of subjects with each genotype classified according to response to 0.4 μm epinephrine (genotype frequency).

A2AR genotype
 G G171 (51.7)32 (58.2)0.654
 G A129 (39.0)19 (34.5)
 A A31 (9.4)4 (7.3)
GNB3 genotype
 C C108 (32.6)10 (18.2)0.03
 T C144 (43.5)24 (43.6)
 T T79 (23.9)21 (38.2)

Discussion

In this study, we have shown that individuals with platelet hyperreactivity to epinephrine also demonstrate increased platelet function when studied using a variety of alternate platelet stimuli (e.g. applied shear stress, other platelet agonists, no stimulus), assay systems (e.g. PFA-100, flow cytometry, aggregometry, Coulter counter), specimen sources (PRP and whole blood) and anticoagulants (citrate and PPACK). The consistent results we observed using these different methods show that platelet hyperreactivity is not specific to epinephrine-mediated aggregation, but rather generalizes to multiple forms of platelet stimulation and even to each major phase of platelet function, from adhesion (increased agglutination in response to low dose ristocetin) to activation (increased P-selectin expression after granule release) to aggregation (increased aggregation in response to multiple different stimuli). Our findings suggest an underlying mechanism that occurs relatively early in the progression of platelet activity (i.e. in proximity to epinephrine's interaction with the α2A-adrenergic receptor) and that is shared by multiple pathways affecting distinct aspects of platelet function.

The consistency of our findings using multiple techniques supports the existence of a true hyperreactive platelet phenotype that cannot be ascribed to mere laboratory artifact. For example, the association between hyperreactivity to epinephrine using citrated specimens and that using specimens anticoagulated with PPACK suggests that the hyperreactive phenotype is independent of both calcium concentration and residual thrombin activity. Although it has been reported that other ‘true’ platelet agonists (such as thrombin) are required to initiate aggregation and that epinephrine merely potentiates this process [18], our data support that at low concentrations of epinephrine, an unusually robust aggregation response can occur despite the absence of thrombin and other added agonists. Furthermore, while higher fibrinogen levels likely contribute to increased aggregation to epinephrine [10], the persistence of differences in platelet function between groups even after adjusting for fibrinogen levels supports that hyperreactivity to epinephrine is an intrinsic platelet property (as does the finding that one-third of the group hyperreactive to epinephrine had a fibrinogen level less than the mean fibrinogen level for the non-hyperreactive group – data not shown). Moreover, the finding that platelet hyperreactivity in PRP is strongly associated with shorter PFA-100 closure times, increased P-selectin expression and increased PAC-1 binding in whole blood suggests that the platelet hyperreactivity observed in PRP also extends to the whole blood milieu (in vitro).

Although the exact mechanisms by which stimulation with epinephrine promotes platelet aggregation remain unclear [19], epinephrine binding to the A2AR on the platelet surface is thought to mediate the platelet response through G-protein activation and subsequent intraplatelet signaling [20,21]. Given the marked variability between healthy individuals’ responses to low concentrations of epinephrine, we hypothesized that variations in genes encoding proteins relevant to this pathway could contribute to the hyperreactive phenotype. The A2AR is a G-protein coupled receptor and as such, activates heterotrimeric G-proteins, including those containing the β3 subunit [22]. The C825T polymorphism located in exon 10 of the gene encoding this subunit (GNB3) is associated with alternative splicing variants which appear to enhance G-protein signal transduction [23,24]. In addition to associations with risk for hypertension, obesity and atherosclerosis [24], the 825T allele has been reported to be associated with enhanced platelet aggregation after epinephrine stimulation [25], although contradicting results were reported when subjects were studied at higher epinephrine concentrations [26]. In our much larger study, the 825T allele was associated with the epinephrine hyperreactive phenotype, providing substantial evidence that this genetic variation may indeed be contributory. However, given the modest strength of the association [OR = 2.9 comparing TT with CC homozygotes, 95% CI = (1.3, 6.4)] and that the GNB3 polymorphism was not associated with increased platelet function according to any of the other assays we performed (data not shown), other factors likely play a role as well. Further study of factors related to A2AR, including direct measurement of A2AR receptor levels, are required so that their relationship to platelet hyperreactivity to epinephrine can be determined.

Interestingly, the agonist ADP accounted for the strongest association with hyperreactivity to epinephrine (adjusted OR = 23.0 with aggregometry) as well as significant associations using other assay systems (flow cytometry and PFA-100; Figs 3 and 4). Although epinephrine and ADP bind to distinct G-protein coupled receptors on the platelet surface, both receptors are thought to couple with Gi family members (thereby inhibiting adenylyl cyclase) [20] and correlation between aggregation to these two agonists has been reported [27]. Future studies should explore potential underlying mechanisms shared by the diverse platelet stimuli and phases of platelet activity highlighted by our findings, including regulation of intraplatelet calcium and cyclic AMP concentrations. The contribution of platelet hyperreactivity to more ‘downstream’ components of platelet function such as platelet procoagulant activity, coated-platelet formation and thrombin generation should also be explored.

One effector molecule critical to platelet aggregation, regardless of stimulus, is the platelet fibrinogen receptor, GPIIb-IIIa. We found that expression of both quiescent and activated forms of this receptor was increased in subjects with platelet hyperreactivity to epinephrine. Although GPIIb-IIIa is the most abundant receptor on the platelet surface, little has been published regarding the impact of variation in expression levels of this receptor on platelet function [28,29]. However, as this molecule serves as the final effector for platelet aggregation, it is not difficult to imagine that higher levels of GPIIb-IIIa expression and activation could contribute to the hyperreactive platelet phenotype, perhaps through enhanced outside-in signaling after fibrinogen binding. In our study cohort, variation in GPIIb-IIIa expression between individuals may have been related to differences in platelet size (r2 = 0.24 for correlation between MPV and GPIIb-IIIa surface expression). However, augmented GPIIb-IIIa expression as a result of larger platelet size alone may be sufficient to confer some degree of platelet hyperreactivity; high MPV has been linked prospectively to increased cardiovascular risk [30,31]. Importantly, increased PAC-1 binding (after ADP stimulation) in the hyperreactive group (Fig. 5) persisted even after adjustment for differences in MPV, supporting that additional factors contribute to this generalized and genuine hyperreactive platelet phenotype.

Validation of our results in other settings is required; well-designed prospective studies directed at assessing the utility of the epinephrine aggregation assay as a measure of an individual's cardiovascular risk are also of utmost importance. This will require a substantial effort, using either a large cohort of high-risk patients or an even larger cohort of healthy subjects who would be followed for incident events. Given the multiple clinical studies that have linked platelet hyperreactivity (using other functional assays) to cardiovascular risk [2–7,30], and the associations described herein between the epinephrine aggregation assay and some of these previously reported measures of platelet reactivity (e.g. MPV [30,31], spontaneous platelet aggregation [2] and P-selectin expression [5]), we anticipate that this assay will prove clinically useful. Recent work in animal models implicating platelet activation in the pathogenesis of atherosclerosis [32,33] and inflammation [34], further highlights the need for a reliable, standardized assay of platelet hyperreactivity; such a tool could lead to more accurate risk assessment for these chronic and highly prevalent disorders. The wide availability of the epinephrine aggregation assay coupled with its impressive reproducibility [10] and physiologic relevance underscore its potential utility in clinical and research settings as a standardized approach for identifying platelet hyperreactivity.

Acknowledgements

This work was supported by grants from the National Institutes of Health (RR17665 and HL81539 to D.L.Y. and HL65229 to P.F.B.) and the National Hemophilia Foundation (D.L.Y.) and Baylor College of Medicine. The authors thank Dr O'Brian Smith for providing consultation on statistical methods, Jennifer Wood and David Lopez for assisting with data acquisition, and Susan Larrucea for synthesizing collagen-related peptide.

Disclosure of Conflict of Interests

The authors state that they have no conflict of interests.

Ancillary