Small-size platelet microparticles trigger platelet and monocyte functionality and modulate thrombogenesis via P-selectin

Authors

  • Silvia Montoro-García,

    1. University of Birmingham Centre for Cardiovascular Sciences, City Hospital, Birmingham, UK
    2. Department of Cardiology, Hospital Clínico Universitario Virgen de la Arrixaca, Instituto Murciano de Investigación Biosanitaria (IMIB), Murcia, Spain
    Search for more papers by this author
  • Eduard Shantsila,

    1. University of Birmingham Centre for Cardiovascular Sciences, City Hospital, Birmingham, UK
    Search for more papers by this author
  • Diana Hernández-Romero,

    1. Department of Cardiology, Hospital Clínico Universitario Virgen de la Arrixaca, Instituto Murciano de Investigación Biosanitaria (IMIB), Murcia, Spain
    Search for more papers by this author
  • Eva Jover,

    1. Department of Cardiology, Hospital Clínico Universitario Virgen de la Arrixaca, Instituto Murciano de Investigación Biosanitaria (IMIB), Murcia, Spain
    Search for more papers by this author
  • Mariano Valdés,

    1. Department of Cardiology, Hospital Clínico Universitario Virgen de la Arrixaca, Instituto Murciano de Investigación Biosanitaria (IMIB), Murcia, Spain
    Search for more papers by this author
  • Francisco Marín,

    1. Department of Cardiology, Hospital Clínico Universitario Virgen de la Arrixaca, Instituto Murciano de Investigación Biosanitaria (IMIB), Murcia, Spain
    Search for more papers by this author
  • Gregory Y. H. Lip

    Corresponding author
    1. University of Birmingham Centre for Cardiovascular Sciences, City Hospital, Birmingham, UK
    • Correspondence: Professor Gregory Y. H. Lip, University of Birmingham Centre for Cardiovascular Sciences, City Hospital, Birmingham B18 7QH, UK.

      E-mail: g.y.h.lip@bham.ac.uk

    Search for more papers by this author

Summary

This study aimed to examine the mechanisms of cellular activation by small-size platelet microparticles (sPMP) and to present the performance of high-resolution flow cytometry for the analysis of subcellular entities from different origins. Plasma counts of sPMP were analysed in coronary artery disease patients (n = 40) and healthy controls (n = 40). The effect of sPMP and platelet debris (PD) in pathophysiologically relevant doses on platelet and monocyte activation parameters and thrombogenesis was investigated via flow cytometry and thromboelastometry. New generation flow cytometry identifies differences in size, levels and surface molecules of sPMP derived in the absence of stimulus, thrombin activation and platelet disruption. Addition of sPMP resulted in platelet degranulation and P-selectin redistribution to the membrane (P = 0·019) in a dose and time-dependent manner. Blood clotting time decreased after addition of sPMP (P = 0·005), but was not affected by PD. Blocking P-selectin (CD62P) in sPMP markedly reverted the effect on thrombus kinetics (P = 0·035). Exposure to sPMP stimulated monocyte expression of intercellular adhesion molecule-1 (P < 0·03) and decreased monocyte interleukin-6 receptor density (P < 0·01). These results implicate sPMP as a direct source of downstream platelet and monocyte activation. In pathological coronary artery disease conditions, higher levels of sPMP favour a prothrombotic state, partly through P-selectin expression.

Blood cells are known to shed biologically active extracellular vesicles, initially termed ‘platelet-like activity of serum’ (O'Brien, 1955). Later, they were named ‘platelet dust’, although their cellular origin and stimuli regulating their release remained unclear (Wolf, 1967; Montoro-García et al, 2011). To date, generation of cell microparticles (MP) is considered to be primarily a consequence of cellular activation and apoptosis but it also occurs in resting platelets without any signs of activation (George et al, 1986; Cauwenberghs et al, 2006). By consensus, MP are defined as small particles (0·1–1 μm diameter) that commonly present phosphatidylserine (PS) in their outer membrane (Freyssinet & Toti, 2010). The discrimination of MP remains technically challenging but high-resolution flow cytometry may overcome this problem (Shet, 2008; Lacroix et al, 2010; Robert et al, 2012). Our group has recently developed a flow cytometry-based technique for discrimination of small-size microparticles (sMPS) smaller than <0·5 μm using polystyrene beads and <0·88 μm according to silica beads (Montoro-García et al, 2012). However, size alone may not be sufficient for the differentiation of different types of extracellular vesicles, and their morphological and phenotype characterization needs to be considered.

Platelet MP (PMP) have attracted much interest (Italiano et al, 2010; Owens & Mackman, 2011) and roles in cardiovascular disease have been proposed (Zahra et al, 2011; Montoro-García et al, 2013). Although platelets may not contribute directly to plaque formation, it is of paramount importance to decipher the mechanisms underlying functional alterations of target cells exposed to PMP (Horn et al, 2012). For instance, kinetic features of whole blood clots in the presence of increased levels of PMP have been demonstrated using thromboelastography (TE) (Suades et al, 2012). No standardized method exists for their measurement, and only a few studies have considered biological MP or silica beads in order to calibrate size gates on flow cytometry. Whilst conventional equipment provides information on PMP sized 0·5–1·0 μm (as estimated using plastic beads), the smaller PMP were beyond the scope of detection in previous studies.

In the present study, we aimed to characterize the morphological and phenotypical differences between small-size PMP (sPMP) released from resting platelets, thrombin activation and platelet disruption [termed as platelet debris (PD)] using our novel technical approach. In addition, we aimed to assess the prothrombotic properties of sPMP by evaluating their ability to induce thrombus formation, platelet and monocyte activation ex vivo.

Study design and subjects

Study participants

In order to determine a pathophysiological relevant dose of plasma sPMP, 40 healthy controls (HC) were compared to 40 age- and sex-matched subjects with stable coronary artery disease (CAD) (i.e. no acute exacerbations or hospital admissions for ≥3 months). CAD was confirmed by coronary angiography. Exclusion criteria were the presence of infectious diseases, chronic inflammatory disorders, malignancy, haemodynamically significant valvular heart disease, atrial fibrillation, renal failure (serum creatinine ≥130 μmol/l) and hormone replacement therapy.

Blood samples were collected in commercial 2 ml sodium citrate (3·2%) Vacutainer® (Becton, Dickinson and Company, Oxford, UK) tubes using a 20-gauge needle (the first 1 ml was discarded) and centrifuged immediately in accordance with the International Society of Thrombosis and Haemostasis (ISTH) recommendations (Lacroix et al, 2010) (Supporting information). The participants were recruited in Sandwell and West Birmingham Hospitals NHS Trust between October 2009 and November 2010. The study was performed in accordance with the Helsinki declaration and approved by Local Research Ethics Committee. All participants provided written informed consent.

Platelet sMP and PD production

Residual platelets (30 × 106 platelets/ml) (after resting, thrombin treatment or disruption) were separated from the supernatant by gentle centrifugation (1000 g, 15 min) and filtration through 1·2 μm pore diameter filters. Small-size PMP and PD were diluted in phosphate-buffered saline (PBS, pH 7·4; Gibco–BRL, Paisley, UK) and isolated from the supernatant by high-speed centrifugation (90 min, 64 000 g, 4°C) in an Allegra 64R centrifuge (Rotor F1010, k-factor 575; Beckman Coulter Inc., Oakley Court, UK) (Fig 1B,C). The supernatant was carefully discarded. The remaining pellet was resuspended in 5 ml filtered PBS and centrifuged again in order to remove traces of thrombin and detergent (Figures S1, S2). The complete removal of platelets was verified at this point by flow cytometry (Fig 1A versus B–C). The final sPMP and PD-enriched suspensions (200 μl) were analysed by high-resolution flow cytometry in order to verify the removal of platelets and freshly used in further functional studies (Fig 1).

Figure 1.

Detection of platelets, sPMP and PD by the Apogee A50. (A and D) Washed platelet concentrate (28 × 103 platelets/μl) plus sPMP (vesicles) released without stimulation before high-speed centrifugation. (D) Staining with annexin V-AF. (B and E) Isolated sPMP released after thrombin stimulation (1 u/ml) and high-speed centrifugation. Residual platelets were completely removed. (E) Staining with annexin V-AF. (C and F) Isolated PD after platelet disruption with 0·75% Tween 20 and high-speed centrifugation. (F) Staining with annexin V-AF. Filtered PBS was used as negative control, containing <5% events in the ‘VESICLE gate’ and lacking of any AnV+ events. AnV binding was performed with 5 mmol/l CaCl2 binding buffer, no AnV+ events were found in the presence of EDTA. ‘VESICLE gate’ was set by exclusion of platelets/apparatus noise and according to polystyrene beads size (‘110 nm’, ‘200 nm’, ‘500 nm’ and ‘1 μm’) and silica beads (‘300 nm’ and ‘880 nm’). No overlap between platelets and subcellular microparticles was found. Cytometer flow rates were adjusted to keep the counting rate less than 5000 events/s. sPMP, small-size platelet microparticles; PD, platelet debris; PBS, phosphate-buffered saline; AF, Axela Fluor; PTL, plaetlets.

Transmission electron microscopy (TEM)

For TEM, pellets of sPMP and PD were washed twice in filtered PBS with the use of Amicon 10 kDa filters (Millipore, Watford, UK). One drop was adsorbed onto a carbon-coated formvar film, which is attached to a metal specimen grid. After 15 min sedimentation, the excess sample was blotted off and the grid covered with a small drop of uranyl acetate 1%, which was blotted off after 30 s. The sample was dried and examined by TEM (Jeol, Peabody, MA, USA) operated at 80 kV.

Labelling of sPMP and PD

Annexin V (AnV) staining was performed in sPMP- and PD-enriched suspensions diluted 1:2 with filtered annexin binding buffer 2× (20 mmol/l HEPES, 0·28 mol/l NaCl, 5 mmol/l CaCl2, pH 7·4) (Fig 1). Five microlitre of annexin V conjugated to Alexa Fluor (AF) 488 nm were added to the dilution, incubated for 30 min, diluted (final dilution 1:10) and processed according to the manufacturer's instructions (Molecular Probes; Invitrogen, Carlsbad, CA, USA). Control samples were prepared in a similar manner with annexin V binding buffer 2× containing 5 mmol/l EDTA instead of CaCl2 (data not shown).

Small-size PMP and PD were labelled for 30 min with 0·05 μg of biotinylated anti-human CD42b [glycoprotein (GP) Iβ, clone AK2; Abcam, Cambridge, UK], anti-human CD61 (GP IIIα, clone PM6/13; MyBiosource, San Diego, CA, USA), anti-human CD62P (P-selectin, clone AK4; Biolegend, San Diego, CA, USA) or anti-human CD63 (granulophysin, clone H5C6; Biolegend) antibodies, followed by a second incubation with 0·25 μg of streptavidin-AF 647 nm-R-phycoerythrin (PE) tandem conjugate (Molecular Probes; Invitrogen) for 20 min and then diluted with 450 μl filtered PBS (final dilution 1:10) (Figure S3). Aliquots of frozen antibodies were stored until use at −20°C. The ratio for biotin-streptavidin concentration (1:5) was optimized in preliminary experiments to avoid a washing step (Montoro-García et al, 2012). Background fluorescence was compared to biotinylated anti-immunoglobulin G1 (anti-IgG1) antibody (0·05 μg) (Abcam and Biolegend) bound to streptavidin-AF 647 nm-R-PE conjugate (0·25 μg).

Flow cytometry

High-resolution flow cytometry

The usefulness of high-resolution flow cytometry to discriminate between sPMP (from resting and stimulated platelets) and artificial PD was evaluated. A gating strategy similar to that used for sPMP was applied to PD using the Apogee A50 flow cytometer (Apogee Flow Systems, Hertfordshire, UK) as previously described (Montoro-García et al, 2012, 2013) (Fig 1). Briefly, polystyrene beads of 110, 200, 500 nm and 1 μm diameter together with 300 and 880 nm silica beads (Apogee Flow Systems) were used to set up the vesicle gate in the two small angle light scatter (LS) detectors (LS1 and LS2) (Fig 1A–C). The voltage for LS1 and LS2 was set at 290V and 470V, respectively. Logarithmic scales were used for all channels. There was overlap in the 110 nm beads population and the instrument noise (determinate with filtered PBS); the threshold was set at 42 and 51 units for LS1 and LS2, respectively. Platelet counts (events/μl) were determined with high-resolution flow cytometry in a LS1–LS2 gate delimited by 0·5–1 μm polystyrene size beads and silica beads larger than 880 nm (Fig 1A). The flow rate and the location of the ‘VESICLE’ gate in the LS1–LS2 datagram were regularly confirmed and remained stable (intra-assay coefficient of variation <10%). Small-size PMP (resting and thrombin) and PD were gated to test their binding to annexin V (Fig 1D–F), anti-CD42b, anti-CD61, anti-CD62P and anti-CD63 in order to assess their morphologically and phenotypically features and compare them to sPMP release without stimulation.

In vitro assessment of platelet activation

Freshly collected citrated blood (450 μl) from a single healthy donor was incubated with 50 μl stock solution sPMP or PD (final concentration at 10 000 and 20 000 events/μl) (n = 6 healthy volunteers for 30 min at room temperature). Surface expression of membrane constitutive CD42a (GPIX) and platelet activation markers (P-selectin and PAC-1 binding) were then assessed as already published (Blann et al, 2013). Before and after incubation, the blood (5 μl) was labelled with 0·05 μg anti-CD42a-peridinin chlorophyll (PerCP) (clone Beb1), anti-CD62P-allophycocyanin (APC) (clone AK-4), PAC-1-fluorescein isothiocyanate (FITC) (clone PAC-1) (which recognizes the conformationally activated GPIIβ/IIIα) (all from BD Biosciences, Oxford, UK) for 15 min, diluted with 1 ml filtered PBS and then analysed in a FACSCalibur flow cytometer (Becton, Dickinson and Company, Oxford, UK). Positive controls of platelet activation were assessed after incubation of the whole blood for 2 min with ADP (0·02 mmol/l; Biodata Corporation, Horsham, PA, USA) and arachidonic acid (0·5 mg/ml; Biodata Corporation). Negative control was also assessed after incubation with the integrin-blocking peptide RGDS (0·075 mmol/l; Sigma Aldrich, St. Louis, MO, USA) for 15 min. Isotype controls were performed with different monoclonal anti-IgG1 (BD Biosciences).

In vitro assessment of monocyte activation

Freshly collected EDTA blood (450 μl) from a single healthy donor was incubated with 50 μl stock solution sPMP or PD (final concentration 100 000 and 300 000 events/μl) from 13 healthy volunteers for 12 h at 37°C. Mouse anti-human monoclonal fluorochrome-conjugated antibodies against CD16-AF 488 (clone DJ130c; AbDSerotec, Oxford, UK), CD14-PE (clone MфP9; BD Biosciences) were mixed in different tubes with 100 μl incubated blood. PE-conjugated antibodies against intercellular cell adhesion molecule-1 receptor [ICAM-1R; (R&D Systems, Europe Ltd., Abingdon, UK) clone 166623] and CXCR4 (clone 12G5; R&D); APC-conjugated antibodies against interleukin-6 receptor (IL6R, clone 17506; R&D) and CD163 (clone 215927; R&D). After incubation for 15 min, red blood cells were lysed by 2 ml of lysing solution® (BD Biosciences) for 15 min, followed by dilution in 1·5 ml of PBS and immediate flow cytometric analysis. Expression of the surface markers was quantified as mean fluorescence intensity (MFI) on different monocyte subsets [CD14++CD16 (Mon1), CD14++CD16+CCR2+ (Mon2), and CD14+CD16++CCR2 (Mon3)] were assessed in a BD FACSCalibur flow cytometer (Becton Dickinson) as previously described (Shantsila et al, 2011; Tapp et al, 2013).

Thromboelastometry (TE)

Thrombus formation and strength were assessed in a thromboelastometric system (ROTEM® delta; Pentapharm GmbH, Münich, Germany). Citrated whole blood (450 μl) from a single healthy volunteer was incubated with 50 μl stock solution sPMP and PD (final concentration of 10 000 and 20 000 events/μl) from 13 healthy donors for 30 min at room temperature. Control experiments were performed with filtered PBS instead of sPMP or PD. EXTEM-S, INTEM-S and FIBTEM-S tests were performed according to the manufacturer's instructions, at 37°C for 35 min. The tests were repeated once in the same machine within 2 h of blood extraction (intra-assay coefficient of variation <7%). Different thromboelastometric parameters were recorded for each test, clotting time (CT), clot formation time (CFT), alpha-angle (α), A10 and A20 value (Amplitude after 10 and 20 min respectively) and maximum clot firmness (MCF).

We next investigated whether P-selectin expressed in sPMP was directly involved in thrombus kinetics. Isolated sPMP (stock at 300 000 events/μl) from 12 healthy donors were incubated with 1 μg anti-CD62P for in vitro blocking (LEAF purified antibodies; Biolegend UK Ltd., Cambridge, UK) or only PBS for 1 h at room temperature. An additional control sample was included with IgG1 isotype antibodies (LEAF purified antibodies; Biolegend UK Ltd.) (n = 3). Samples were diluted with 5 ml filtered PBS and then high-speed centrifuged again to remove traces of antibodies. Fresh citrated whole blood (450 μl) from the one healthy volunteer was diluted with those three solutions (50 μl) up to 20 000 events/μl and incubated for 30 min. Finally, EXTEM-S test was assessed with the four following blood preparations: PBS, sPMP [unblocked sPMP] and sPMP with anti-CD62P [blocked sPMP] and isotype control.

Statistical analysis

Continuous variables were tested for normality of distribution with the Kolmogorov–Smirnov test. Normal data are presented as mean ± standard deviation (SD) and non-normal data are presented as median (interquartile range). Comparisons were performed using analysis of variance (anova) with Tukey's post-hoc analysis (normally distributed data) or Mann–Whitney U test (non-normally distributed data). Correlations were tested by Pearson (normally distributed data) or Spearman's rank (non-normally distributed data) correlation method. The analyses were done using spss 19.0 for Windows software (SPSS Inc., Chicago, IL, USA). A two-tailed probability value of P < 0·05 was considered significant in all statistical analyses.

Results

Morphological characteristics of sPMP and PD

Small-size PMP produced in resting conditions and after thrombin stimulation appeared in the gate delimited by 880 nm silica beads below platelets (Fig 1A–B). PD also appeared in the same vesicle gate, mostly below 300 nm silica beads (Fig 1C).

From identical platelet counts and preparation, higher small-size annexin V binding microparticles (sAMP) counts were released under thrombin stimulation compared to resting conditions (29·8 ± 21·0 vs. 4·2 ± 1·0 105/μl, P < 0·05) (Table 1). CD42b+, CD61+ and CD63+ counts were not altered after thrombin stimulation. The major phenotypic difference between thrombin and resting conditions was the lower CD42b expression (342 ± 3·5 vs. 350 ± 3·8 MFI, P < 0·05) and that they are more CD62P+ (13·1 ± 11·9 vs. 2·0 ± 1·6 104/μl, P < 0·01).

Table 1. Visual flow cytometric differences between sPMP and PD (Apogee A50).
 Resting conditionsThrombin stimulationTween 20 disruptionP-value
  1. PD, platelet debris; sPMP, small-size platelet microparticles; AnV, annexin V; LS1, Light scatter; FL1 and 3, Fluorescent channel 1 and 3, respectively.

  2. Counts of sPMP and PD produced from equal platelets concentrations. Results from seven independent experiments for the three conditions are summarized. Resting conditions were assessed in Tyrode's salt (5 mmol/l, CaCl2) without any treatment for 30 min at room temperature. Data are expressed as mean ± standard deviation. anova with Tukey's Post hoc analysis.

  3. Samples were processed per triplicate by high-resolution flow cytometry.

  4. a

    P < 0·05 vs. resting.

  5. b

    P < 0·01 vs. resting.

  6. c

    P < 0·05 vs. thrombin.

  7. d

    P < 0·01 vs. thrombin.

AnV+ counts (105/μl)4·2 ± 1·029·8 ± 21·0a44·2 ± 27·4b0·002
LS1 mean217·0 ± 18·7225·0 ± 31·8231·0 ± 27·70·64
FL1 mean AnV+226·0 ± 21·2225·0 ± 15·0196·0 ± 20·0a,c0·012
CD42b+ counts (104/μl)3·15 ± 2·98·0 ± 5·415·1 ± 3·5b,c0·001
LS1 mean280·0 ± 21·7286·0 ± 8·0326·0 ± 12·6b,c0·005
FL3 mean CD42b+350·0 ± 3·8342·0 ± 3·5a337·0 ± 6·1b0·011
CD61+ counts (103/μl)4·8 ± 1·54·9 ± 1·97·1 ± 5·30·32
LS1 mean297·0 ± 39·2293·0 ± 15·1312·0 ± 22·20·53
FL3 mean CD61+385·0 ± 47·8361·0 ± 11·6350·0 ± 3·80·16
CD62P+ counts (104/μl)2·0 ± 1·613·1 ± 11·9b3·4 ± 1·70·03
LS1 mean279·0 ± 14·9272·0 ± 16·5319·0 ± 14·5b,d<0·001
FL3 mean CD62P+344·0 ± 9·8340·0 ± 1·3336·0 ± 4·30·11
CD63+ counts (103/μl)2·3 ± 0·93·4 ± 0·92·7 ± 1·40·25
LS1 mean275·0 ± 21·3279·0 ± 19·1288·0 ± 22·20·52
FL3 mean CD63+354·0 ± 10·6358·3 ± 7·3350·0 ± 11·30·25

AnV binding, together with anti-CD42b and anti-CD62P antibodies, enabled differentiation between sPMP and PD. Platelet disruption also lead to increased AnV+ counts (P < 0·01 vs. resting conditions) but with lower expression of PS in the membrane compared to sPMP (P = 0·012). Platelets and sPMP after thrombin stimulation did express CD62P (P = 0·03, Figure S3B). PD and residual platelets after disruption and resting did not express CD62P (Table 1, Figure S3A, C). Using the LS1 scale as a proven reference for size, PD seemed ‘larger’ than sPMP (LS1 MFI CD42b+ P = 0·005; LS1 MFI CD62P+ P < 0·001), probably because PD includes micelles of detergent (Figs 1 and 2). The study of the morphological features was assessed with negative staining in TEM (Fig 2). Apart from the size, the negative staining also gives a clear picture of homogenous membrane structure of sPMP compared to larger, high-density particles that are PD.

Figure 2.

Transmission electron micrograph of sPMP (A–C) and PD (D–F). (A, D) Undiluted samples. (B–C, E–F) diluted 1/10 in PBS buffer. While sPMP are spherical and generally ≤200 nm, PD are irregular structures of around 500 nm diameter. Black dots indicate uranyl precipitates. sPMP, small-size platelet microparticles; PD, platelet debris; PBS, phosphate-buffered saline.

AnV affinity to sPMP is probably higher than anti-GP antibodies becuase detected sAMP counts were 10 times higher than CD42b+ and CD62P+ sPMP and 100 times higher than CD61+ and CD63+ sPMP counts. Thus, sPMP were quantified according to their AnV binding (sAMP) in subsequent functional studies.

sAMP counts in coronary artery disease

Patients with CAD had twofold higher plasma sAMP counts compared to sex- and age-matched HC [22 (14–38) × 103 sAMP/μl vs. 13 (8–23) × 103 sAMP/μl, P = 0·003) (Table 2). In order to simulate a pathophysiological state, sPMP and PD were added to whole blood at final concentration of 10–20 × 103 AnV+ events/μl in the following in vitro studies. Such doses (10–20 × 103 AnV+ events/μl) were close to the pathological levels expected in whole blood.

Table 2. Demographic data and plasma sAMP counts in respect to the pathophysiological status.
 CAD (n = 40)HC (n = 40) P
  1. CAD, coronary artery disease patients; HC, healthy controls; sAMP, small-size AnV+ binding microparticles.

Age, years60 ± 1159 ± 130·92
Male, n [%]33 [82]32 [80]0·99
103 sAMP/μl22·3 [13·7–38·0]13·4 [7·5–22·9]0·003

Platelet interaction with sPMP increases P-selectin expression but not PAC-1 binding

Addition of 10 000 sPMP/μl was sufficient to produce phenotypical changes in platelets. After incubation, CD62P (P-selectin) expression increased in platelets in a time-dependent manner (P = 0·019) (Table 3a). Platelet phenotype remained unaltered in the presence of PBS or PD (P = 0·82 and P = 0·60, respectively). Addition of sPMP had no effect in PAC-1 binding (Table 3b). This effect was dose-dependent as 20,000 sPMP/μl produced faster CD62P expression (8·5 ± 2·7 in 15 min, data not shown).

Table 3. Effect of sPMP pathophysiological relevant levels in platelet activation. Citrated whole blood (450 μl) was incubated with sPMP and PD (10 000 events/μl) or PBS (n = 6).
 Time courseP- value
t = 0 mint = 15 mint = 30 min
  1. Results from six independent experiments for the three time points are summarized. Samples were processed per triplicate by FCM (FACS Calibur).

  2. Data are expressed as mean ± standard deviation or median [interquartile range]. anova with Tukey's Post hoc analysis (after ln transformation if the case). sPMP, small-size platelet microparticles; PD, platelet debris; PBS, phosphate-buffered saline.

  3. a

    P < 0·05 vs. Control.

  4. b

    P = 0·067 vs. PD.

  5. c

    P = 0·058 vs. PD.

(a) P-selectin (CD62P) expression in platelets
Control with PBS5·2 ± 3·54·2 ± 2·54·8 ± 2·00·82
sPMP4·2 ± 2·57·8 ± 2·4b8·8 ± 2·8a,c0·019
PD3·9 ± 2·74·5 ± 2·15·4 ± 1·90·60
 P-value0·790·0330·02 
(b) PAC-1 binding in platelets
Control with PBS1·1 [0·4–2·0]0·3 [0·2–2·4]0·2 [0·05–3·1]0·45
sPMP0·3 [0·2–2·4]0·4 [0·2–2·9]0·3 [0·03–3·6]0·72
PD0·2 [0·2–3·0]0·3 [0·2–4·3]0·3 [0·01–3·9]0·73
 P-value0·640·730·81 

Whole blood interaction with sPMP decreases clotting time

The addition of sPMP (20 000 sPMP/μl) significantly reduced clotting time in the EXTEM-S test (CT EXTEM) (by >15%), while PD did not change any thrombus kinetics or strength (Table 4a). The FIBTEM-S test includes cytochalasin D, a platelet inhibitor; addition of sPMP or PD produced no significant effect in this test (Table 4b). Similarly, the INTEM-S test showed no significant differences after the addition of sPMP and PD (data not shown). Lower sPMP counts (10 000 sPMP/μl) did not produce quantifiable results with TE (data not shown).

Table 4. Effect of sPMP pathophysiological relevant levels in thrombogenesis kinetics. Citrated whole blood (450 μl) was incubated with sPMP and PD (20 000 events/μl) or 50 μl PBS. (a) Thrombus formation and strength parameters after incubation, EXTEM-S test. (b) Thrombus formation and strength parameters after incubation, FIBTEM-S test.
 CTCFTAlpha (α)A20MCF
(a)
Control with PBS70·2 ± 10·2148·3 ± 48·562·9 ± 8·152·1 ± 10·254·1 ± 5·0
sPMP57·5 ± 12·0a136·6 ± 29·165·3 ± 6·653·1 ± 3·555·0 ± 4·4
PD65·7 ± 9·7150·1 ± 27·863·0 ± 5·452·2 ± 4·154·5 ± 3·7
P-value0·0050·640·590·820·85
 CTA10A20MCF
  1. Results from 13 (sPMP) and 11 (PD) independent experiments are summarized. Samples were processed per duplicate by thromboelastometry.

  2. Data are expressed as mean ± standard deviation or median [interquartile range]. anova with Tukey's Post hoc analysis (after ln transformation if the case).

  3. sPMP, small-size platelet microparticles; PD, platelet debris; PBS, phosphate-buffered saline; CT, clotting time; CFT, clot formation time; α, alpha-angle; A20, amplitude after 20 min respectively and MCF, maximum clot firmness.

  4. a

    P = 0·003 vs. control.

(b)
Control with PBS68·4 ± 10·69·0 [8·0–13·0]9·0 [9·0–13·0]9·0 [9·0–13·0]
sPMP63·3 ± 12·78·5 [7·3–9·0]8·5 [8·0–10·0]9·0 [8·3–11·0]
PD61·2 ± 9·59 [8·0–12·0]10·0 [8·0–11·0]10·0 [9·0–12·0]
P-value0·160·280·210·38

Thrombin activation is CD62P+ sPMP dependent

Cytochalasin D conceals the sPMP effect on thrombogenesis kinetics and subsequently, platelets seem to be directly involved. As sPMP bear higher P-selectin (CD62P) expression and further activate platelets, we decided to block CD62P+ sPMP before blood incubation. No differences in CT were found after blocking CD62P+ sPMP and control, while unblocked-sPMP still decreased CTEXTEM (P = 0·003) (Table 5). No difference was found between isotype control-sPMP and unblocked-sPMP (n = 3, data not shown).

Table 5. Thrombus formation and strength parameters after incubation, EXTEM-S test. Whole blood (450 μl) was incubated for one hour with 50 μl PBS, unblocked sPMP or CD62P blocked-sPMP (20 000 events/μl).
 CTCFTAlpha (α)A20MCF
  1. sPMP, small-size platelet microparticles; PBS, phosphate-buffered saline; CT, clotting time; CFT, clot formation time; α, alpha-angle; A20, amplitude after 20 min respectively and MCF, maximum clot firmness.

  2. Results from twelve sPMP (blocked/unblocked) independent experiments are summarized. Samples were processed per duplicate by thromboelastometry.

  3. anova with Tukey's Post hoc analysis. Data are expressed as mean ± SD.

  4. a

    P = 0·003 vs. control.

Control with PBS54·6 ± 7·8132·9 ± 11·966·3 ± 2·554·8 ± 2·1456·8 ± 2·5
Unblocked-sPMP (Positive control)46·2 ± 7·4a124·9 ± 18·769·7 ± 3·456·0 ± 3·257·8 ± 3·9
CD62P blocked-sPMP52·6 ± 8·5123·3 ± 17·468·3 ± 4·355·9 ± 2·757·6 ± 2·7
P-value0·0350·320·070·460·67

Effect of sPMP and PD on monocyte activation

No significant changes were observed after 6 h incubation of whole blood with 100 000 of sPMP/μl. Small-size PMP counts were increased up to 300 000 sPMP/μl for 12 h and this addition increased ICAM-1R expression in all three monocyte subsets (P = 0·003; P = 0·017 and P = 0·027, respectively) (Table 6). Addition of sPMP significantly decreased IL6R density on Mon1 (P = 0·008) and Mon2 (P = 0·003) (Table 6). No effect of sPMP was observed on monocyte CXCR4 and CD163 expression. The addition of PD did not affect the surface expression of any monocyte marker.

Table 6. Expression of reparative and inflammatory markers in the three subsets of monocytes in response to sPMP addition. EDTA whole blood (450 μl) was incubated with sPMP and PD (300 000 events/μl) or 50 μl PBS for 12 h (n = 13).
 ICAM1R_Mon1ICAM1R_Mon2ICAM1R_Mon3
Control with PBS232·0 ± 17·7226·2 ± 16·562·8 ± 15·7
sPMP269·5 ± 30·2b,c254·8 ± 16·9a85·1 ± 36·3c
PD234·9 ± 36·0244·9 ± 36·159·0 ± 14·6
P-value0·0030·0170·027
 IL6R_Mon1IL6R _Mon2IL6R_Mon3
Control with PBS11·16 ± 3·114·2 ± 3·216·8 ± 4·1
sPMP7·56 ± 3·19·9 ± 2·816·3 ± 6·3
PD14·2 ± 3·211·7 ± 2·616·0 ± 7·5
P value0·0080·0030·158
 CXCR4_Mon1CXCR4_Mon2CXCR4_Mon3
Control with PBS44·5 ± 5·964·1 ± 9·027·0 ± 8·0
sPP44·3 ± 9·264·6 ± 13·722·0 ± 5·2
PD44·1 ± 2·865·0 ± 5·122·9 ± 7·7
P- value0·980·970·18
  1. sPMP, small-size platelet microparticles; PD, platelet debris; PBS, phosphate-buffered saline; Mon1/Mon2/Mon3, monocyte subsets 1, 2 and 3, respectively; ICAM1R, intercellular adhesion molecule receptor; IL6R, interleukin-6 receptor.

  2. Data are expressed as mean ± SD. anova with Tukey's Post hoc analysis.

  3. a

    P < 0·05 vs. control.

  4. b

    P < 0·005 vs. control.

  5. c

    P < 0·05 vs. PD.

 CD163_Mon1CD163 _Mon2CD163 _Mon3
Control with PBS23·2 ± 6·526·5 ± 8·55·8 ± 1·0
sPMP24·8 ± 6·825·6 ± 6·86·1 ± 1·8
PD28·6 ± 6·827·8 ± 10·15·6 ± 1·2
P-value0·200·650·74

Discussion

This study demonstrates for the first time that high-resolution flow cytometry reliably discriminates sPMP from different conditions (resting and stimulated) and PD (Lee et al, 2011). Secondly, the negative staining method (TEM) provided a useful analytical tool for identifying structural features of sPMP and PD.

Small-size PMP produced under resting and thrombin conditions share a unique pattern of reactivity with the workshop antibodies denoting meaningful phenotypic differences. In fact, the reactivity of anti-CD42b was greatest in sPMP from resting platelets and least from thrombin-stimulated platelets, in accordance with our previous report (Montoro-García et al, 2012). In the presence of thrombin, CD42b is internalized whist CD62P is translocated to the outer plasmatic membrane (Hourdille et al, 1990; Ruf & Patscheke, 1995). Nevertheless, sPMP released under thrombin stimulation did not differ in AnV expression, only in counts. Both sPMP types had similar diameters (100–300 nm silica beads). Overall, differential expression of antigens between resting and thrombin stimulation revealed that AnV binding can be used for sPMP quantification, irrespective of pathophysiological status, while P-selectin denotes sPMP released from platelet activation only. According to some studies, platelet activation and apoptosis are indeed separated processes (Gyulkhandanyan et al, 2013; Mutlu et al, 2013). Moreover, binding with AnV was higher compared to anti-GP antibody binding. These results indicate that most sPMP express PS on their outer membrane while protein/GP staining of such small surfaces is insufficient for accurate quantification. On the other hand, sPMP and PD differences were also prominent. Platelet disruption produces the release of higher entities bearing PS but not P-selectin.

In order to estimate the pathophysiological sAMP levels in a prothrombotic state, we studied patients with CAD and compared them to matched healthy subjects. We found that addition of pathophysiological relevant doses of sPMP decreased CT EXTEM and remained unaltered in the presence of PD. Furthermore, the experiment also showed that the effect was dependent on platelets because the kinetics of thrombus formation in FIBTEM-S test (with cytochalasin D) did not change after incubation with sPMP. The coagulation activation is similar in EXTEM-S and FIBTEM-S however, the INTEM-S test does not use tissue factor (TF) but the intrinsic pathway (i.e., Factor VIII). It is not clear how sPMP are involved in the extrinsic pathway; some studies suggest that sMP carry TF on their surface, promoting thrombus formation in vivo (Aleman et al, 2011; Collier et al, 2013). However, our group has been unable to quantify this marker on the sPMP surface with the current approach. In the search for another mechanism that might regulate extrinsic coagulation, PS is also well studied and eventually potentiates the activation of thrombin from prothrombin by activated factor X (Hemker et al, 1983). This process (reflected by a decreased CT) was facilitated by sPMP but not by PD, both bearing PS. Thus, our results suggest that PS alone may not be sufficient to accelerate thrombogenesis.

One of the major contrasts between sPMP released in resting conditions and those produced under thrombin stimulation is a higher CD62P expression. The study shows that tagging CD62P-sPMP leads to the recovery of CT, and that sPMP can induce expression of CD62P in platelets. Interestingly, surface expression of CD62P on activated platelets has been shown to be an irreversible phenomenon, which induces formation of platelet-monocyte aggregates and promotes vascular inflammation and thrombosis (Passacquale et al, 2011; Czepluch et al, 2013). Moreover, P-selectin has also been proposed as a novel therapeutic target in vascular disease (Chelliah et al, 2009; Japp et al, 2013; Kling et al, 2013).

Given that sPMP mediate platelet activation and reduce CT, we evaluated whether sPMP can also modulate monocyte phenotype. Initially, over-pathophysiological doses (100 000 sPMP/μl) of sPMP were insufficient to produce any monocyte impairment, perhaps because monocyte activation and maturation occurs slowly in response to chronic more than acute states. Our data demonstrates that higher counts of sPMP influence monocyte phenotype by increasing ICAM-1R expression on the three different monocyte subsets. In accordance, MP isolated from atherosclerotic plaques have been shown to stimulate endothelial monocyte adhesion in vitro (Rautou et al, 2011). This process is disrupted by anti-ICAM1 neutralizing antibodies, suggesting a link between the addition of sPMP and ICAM1R in monocyte adhesion. Also, human blood monocytes constitutively express IL6R (Bauer et al, 1989). Treatment of cultured monocytes with IL6 results in a decrease of IL6R at the transcriptional level (Bauer et al, 1989). Although it is not clear how, pathological doses of sPMP might also induce the IL6R targeting of soluble IL6. Whether IL6R decreases as a consequence of lower transcription or higher IL6 expression cannot be formally demonstrated here, but those observations fairly support sPMP-dependent changes in monocytes.

Study strengths and limitations

Small-size PMP were not produced under pathophysiological conditions (but thrombin 1u/ml) and this phenotype might thus be impaired in vivo. We also used whole blood from healthy donors -not CAD patients – which limits the extent to which the results can be translated into the clinical setting. Nevertheless, we show that sPMP produced in vitro have the ability to interact with neighbouring platelets and monocytes (and therefore may take part in primary haemostasis). It is important to mention that PD proteins and lipids remain native and active when immersed in the micellar structure which allows the analysis of their membrane composition and if the case, activity (Muhle-Goll et al, 2012). Our ex vivo stimulation/disruption model does not undermine the efficient discrimination of subcellular entities with the flow cytometric approach.

Conclusions

There are significant phenotypic differences between sPMP depending on the stimulus, which should be taken into account for future sPMP quantification. The findings of this study demonstrate that, far from PD, sPMP have important and complex implications as key systems regulating platelets, monocytes and thrombogenesis.

Acknowledgement

We gratefully acknowledge Servicio de Microscopía (SACE) from the University of Murcia for technical support.

Funding

This work was supported by the Heart Research UK [RG 2579/09/10] and a European Society of Cardiology Research Grant. SMG held postdoctoral position from the Instituto de Salud Carlos III, ‘Sara Borrell program’ (Spain). DHR is supported by a postdoctoral contract from ‘Instituto Murciano de Investigación Biosanitaria, IMIB’, Murcia, Spain. Until January 2014, she held a ‘Sara Borrell’ contract from the Instituto de Salud Carlos III. EJ holds a predoctoral grant from the Instituto de Salud Carlos III (Spain).

Competing interests

The authors have no competing interests.

Authorship

SMG and ES performed analysis, analysed data and wrote the paper; DHR and EJ contributed essential reagents and tools. MV and FM corrected the manuscript; GYHL designed the research study and corrected the manuscript.

Ancillary