Kovit Pattanapanyasat, PhD, Office for Research and Development, Department of Immunology, Faculty of Medicine, Office for Research and Development, Siriraj Hospital, Mahidol University, Bangkok 10700, Thailand. E-mail: email@example.com
Thromboembolic complications have been documented in thalassaemia patients. The aggregability of abnormal red blood cells and the high level of membrane-derived microparticles (MPs) stemming from blood cells are thought to be responsible for the associated thrombotic risk. We investigated the number of MPs, their cellular origin and their procoagulant properties in β-thalassaemia. Fresh whole blood was simultaneously stained for annexin V, cellular antigens and the known density beads. The procoagulant properties of these phosphatidylserine (PS)-bearing MPs were also measured by assessing the platelet factor-3-like activity in the blood. Flow cytometric results showed that splenectomised β-thalassaemia/HbE patients had significantly higher levels of PS-bearing MPs than non-splenectomised β-thalassaemia/HbE patients and normal individuals (P < 0·0001). There was a good correlation between PS-bearing MPs and PS-bearing platelets, reflecting the existence of chronic platelet activation in β-thalassaemia/HbE patients (rs = 0·511, P < 0·001). The cellular origin of PS-bearing MPs showed mostly activated-platelet origin with adhesion (CD41a/CD62P/CD36). Moreover, the platelet procoagulant activity was higher in splenectomised β-thalassaemia/HbE patients when compared with non-splenectomised (P < 0·05) and normal individuals (P < 0·01), and the amount correlated with PS-bearing MPs (rs = 0·560, P < 0·001). These findings suggest that MPs originate from activated platelets with a potential to aggravate thrombotic events when the numbers are excessive, as is commonly seen in splenectomised β-thalassaemia/HbE patients.
In the circulation, the presence of high levels of membrane-derived microparticles (MPs), shedding from the plasma membrane of most eukaryotic cells, with a large interest on those derived from platelets and leucocytes, undergoing activation or apoptosis have been shown to enhance procoagulant activity (Hugel et al, 1999; Mallat et al, 1999; Freyssinet, 2003). It is also possible that RBC membrane-derived vesicles may contribute to defects in coagulation (Pattanapanyasat et al, 2004). For example, the RBC membrane undergoes vesiculation under a variety of conditions, which include increased cytoplasmic calcium concentration, reduced ATP content and disruption of the membrane lipid–protein organization (Wagner et al, 1986) – all of which are features characteristic of thalassaemic RBCs. It has been shown that vesicles and MPs from platelets and RBC strongly bind annexin V, a protein known for its interaction with negatively charged phospholipids such as phosphatidylserine (PS) (Shinar et al, 1987; Borenstain-Ben Yashar et al, 1993; Zwaal & Schroit, 1997; Kuypers et al, 1998). PS is one of the key membrane phospholipids that is normally located along the inner side of the membrane bilayer. It is known to activate the alternative complement pathway and to support the formation of activated clotting enzymes when expressed on the outer lipid bilayer (Borenstain-Ben Yashar et al, 1993; Zwaal & Schroit, 1997; Kuypers et al, 1998; Freyssinet, 2003). An important insight into this procoagulant potential was corroborated by clinical studies showing elevated levels of circulating MPs in patients with an increased risk for thromboembolic events (i.e. patients with myocardial infarction, or disseminated intravascular coagulation) (Nieuwland et al, 1997; Combes et al, 1999; VanWijk et al, 2003).
Several observations indicate that RBCs from β-thalassaemia, particularly following splenectomy, show enhanced cohesiveness and an increased procoagulant effect, which seems to be due to the increased expression of PS on the RBC surface (Borenstain-Ben Yashar et al, 1993; Helley et al, 1996). There have been a very few studies characterising MPs, their cellular origin and their relationship with the hypercoagulable state in thalassaemia. We postulated that the hypercoagulable state found in thalassaemia patients is partly linked to the presence of increased amounts of circulating MPs and that these MPs promote membrane-associated procoagulant activities. To test this hypothesis, we first determined if there was any direct evidence for the presence and quantities of circulating MPs in thalassaemia. We next examined the cellular origin of these circulating MPs in these patients. Levels of MPs in the peripheral blood were measured by using a quantitative flow cytometric technique as described (Pattanapanyasat et al, 2004). MPs were identified by their size and the use of monoclonal antibodies (mAb) to determine the cellular origin of the MPs. We also addressed the relationship between these MPs and their procoagulant activity by using a platelet factor-3 assay.
Materials and Methods
Patients and blood samples
Venous blood samples were collected from a total of 91 β-thalassaemia/Hb E patients ranging in age from 18 to 42 years and included 38 patients who had been splenectomised for more than 5 years. The diagnosis of thalassaemia for all subjects was made by standard hematological techniques and Hb analysis. For control purposes, venous blood from 35 otherwise healthy and age-matched individuals was utilised. Blood samples from six idiopathic thrombocytopenic purpura (ITP) patients who have undergone splenectomy served as pathological controls. All subjects had normal glucose-6-phosphate dehydrogenase levels, had no evidence of concurrent infection, and none had been hospitalised or transfused within 8 weeks. There was no history of vaso-occlusive episode or atherosclerosis vascular disease in these patients prior to the initiation of the study or during the sampling period. After obtaining informed consent, 4 ml of venous blood was collected by venepuncture and aliquoted equally into tubes containing trisodium heparin or 3·2% trisodium citrate (Becton Dickinson Biosciences [BDB], San Jose, CA, USA). The heparinised blood samples were used for flow cytometric immunophenotyping and the complete blood count was performed on the citrate-anticoagulated blood with the use of a Sysmex K800 haematology analyzer (Sysmex, Tokyo, Japan). All blood samples were collected at room temperature (24–26°C) and processed within 6 h. This study was approved by the Ethical Committee of the Faculty of Medicine, Siriraj Hospital, Mahidol University. Approval number 74/2004.
Reagents and mAb
The following mAbs were purchased from BDB: mAb to leucocytes (anti-CD45, clone 2D1), monocytes (anti-CD14, clone MØP9), granulocytes (anti-CD66e, clone B6·2/CD66), platelet glycoprotein IIb/IIIa (anti-CD41a, clone HIP8), P-selectin (anti-CD62P, clone AC1·2), E-selectin (CD62E, clone 685HII) and adhesion molecule (anti-CD36, clone CB38). The mAb to CD41a; clone VIPL3 was also purchased from Caltag Laboratories, Burlingame, CA, USA. The mAb to endoglin (anti-CD105, clone 166707) was from Serotec, Oxford, UK. The mAb to tissue factor (TF) was obtained from American Diagnostica Inc., Stamford, CT, USA, and the mAb to RBCs (anti-glycophorin A, clone JC159) was purchased from Dako, Glostrup, Denmark. All of these mAbs were directly conjugated either with FITC or phycoerythrin (PE), peridinin cholophyll protein (PerCP) or tricolor (TC). FITC- and PE-conjugated annexin V was purchased from PharMingen, San Diego, CA, USA. FITC-F(ab’)2 or PE- F(ab’)2 goat anti-mouse IgG was also purchased from BDB and used as isotype control. All mAbs were utilised at the concentration recommended by the manufacturer.
Staining of PS-positive MPs in peripheral blood
Ten microlitres of each heparinised whole blood sample was diluted with HBSS–HEPES buffer (Hanks’ balanced salt solution buffered with 5 mmol/l N-2-hydroxy-ethylpiperazine-N’-2-ethanesulfonic acid) at 1:4 (vol/vol). Five microlitres of this diluted blood was incubated for 30 min at room temperature in the dark with 10 μl FITC-conjugated annexin V in the presence of either 2·5 mmol/l CaCl2 or 2·5 mmol/l EDTA in a total volume of 100 μl adjusted with HBSS–HEPES. Fifty microlitres of the incubation mixture was then pipetted into tubes preloaded with a known density of fluorescent TruCountTM bead lyophilised pellet (BDB) plus 400 μl of HBSS–HEPES buffer and analysed by using a flow cytometer. To avoid any sedimentation of beads, the samples were thoroughly mixed prior to flow cytometric analysis.
Staining of MPs with different cellular origins
To identify the cellular origin of the MPs in the peripheral blood, 2 μl aliquots of heparinised blood sample was simultaneously stained with 2 μl of annexin V-FITC or annexin V-PE and 10 μl of a variety of fluorochrome-conjugated mAbs specific for glycophorin A, CD45, CD14, CD66e, CD41a, CD36, CD62P, CD105, CD62E and TF in TruCountTM tubes followed by the addition of HBSS–HEPES buffer-containing either 2·5 mmol/l CaCl2 or 2·5 mmol/l EDTA to give a final volume of 100 μl. These incubation mixtures were incubated for 30 min at room temperature in the dark. After incubation, the stained sample was diluted with 300 μl of HBSS–HEPES buffer and analysed by flow cytometry within 30 min.
Flow cytometric analysis
Microparticles were analysed by using a FACSCaliburTM (BDB) flow cytometer. The MPs were excited with 488-nm light from a 15-mW argon ion laser. Logarithmic green, orange and red fluorescence of FITC, PE and PerCP or TC were measured through 530/30 band pass, 585/42 band pass and 670 long pass filters, respectively. Instrument fluorescence calibration and sensitivity were performed by using Calibrite beadsTM (BDB). Data from at least 50 000 events were acquired and analysed with the use of cellquestTM software (BDB). MPs, platelets, intact RBCs and the TruCountTM fluorescent beads were identified by their size as assessed by their logarithmic amplification of the forward light scatter (FSC-H) and sideward light scatter (SSC-H) signals. As seen in Fig 1A, the size of the MPs (R1) was small when compared with the size of platelets. The platelet population, RBCs and TruCountTM beads were represented in region R2, R3 and R4, respectively. The percentage of FITC-annexin V+ MPs in region R1was determined from the two-parameter dot plot of SSC-H/-annexinv V+ fluorochrome indicated by region R5 (Fig 1B). The percentage of FITC-annexin V+ MPs expressing a specific surface antigen (i.e. CD41a) in region R5 was determined by the histogram as indicated by M2 gate (Fig 1C). To calculate the absolute number of annexin V+ MPs with CD41a+, the TruCountTM beads in the staining tube were distinguished from other cellular populations (i.e. CD41a and glycophorin A), and gated by using PE and TC dot plot as represented in region R6 (Fig 1D). The TruCountTM fluorescent beads were acquired for at least 2000 events. The absolute number annexin V+ MPs with CD41a antigen was calculated by using the following formula:
Internal processing control
As the whole blood analysis of MPs in this study is new and any kind of laboratory stress might cause platelet vesiculation, determination of the levels of platelets and RBCs following immunostaining with mAb or during the storage period was chosen to monitor the number of MPs. No evidence for any instability of platelets and RBCs within 6 h, as demonstrated by normal flow cytometric FSC-H and SSC-H light scatter patterns was found. There were also no significant changes in the levels of platelets and RBCs in the blood samples during the processing time. Moreover, the platelets counts, identified as CD41a+ events in the region R2 (Fig 1), agree favourably with those obtained from the automated haematology analyzer with a correlation coefficient of rs = 0·885 (data not shown).
Procoagulant activity using platelet factor-3 assay in whole blood
Platelet-derived procoagulant activity has been termed ‘Platelet factor-3’ as assayed by Russell's viper venom. Platelet factor-3 plays a very important role in the activation of coagulation factors and is regarded to be available during activation of platelets. The total activity was measured following the activation by kaolin or ellagic acid. Briefly, 100 μl of 3·2% trisodium citrated whole blood was added to a mixture containing 600 μl of Veronal buffer pH 7·35, 100 μl of 1 × 10−4 mmol/l ellagic acid (Sigma Chemical, St Louis, MO, USA) and 100 μl of 5 × 10−5 mmol/l MD 805 [(2R, 4R) 4-methyl-1-[N alpha-(3-methyl-1,2,3,4-tetra-hydro-8-quinoline-sulfonyl)l-arginyl]-2-piperidine carboxylic acid monohydrate] (Mitsubishi Kasei Co, Yokohama, Japan) for initiation of the activation process. After incubation for 5 min at 37°C, 100 μl of 25 mmol/l CaCl2 (Sigma) was then added and incubated for another 20 min at 37°C. The activation process was stopped by the addition of 100 μl of 25 mmol/l EDTA (Sigma) followed by centrifugation at 500 × g for 3 min. Twenty-five microlitre of the supernatant used as a source of thrombin was added to the mixture solution containing 375 μl of Tris–imidazole buffer pH 8·1 and 100 μl of 2·5 mmol/l thrombin substrate (S-2238, Sigma). After incubation for 10 min at 37°C, 100 μl of acetic acid was added to the mixture to stop the hydrolysis of thrombin substrate by thrombin. The absorbance of the reaction mixture was measured at 405 nm utilising a Shimadzu UV-160 spectrophotometer (Shimadzu, Kyoto, Japan).
All descriptive statistics (mean, SD, median, interquartile range (IQR), coefficient of variation and ranges) were performed by using the Statistical Package for the Social Sciences (spss), version 11.0 (SPSS Inc., Chicago, IL, USA). Comparisons of statistical difference between parameters were performed by utilising the non-parametric Mann–Whitney U-test. The simple linear regression and Spearman's correlation coefficient (rs) were also determined. For MP values among the groups of patients, the threshold for statistical significance for all comparisons was chosen as P < 0·05.
Haematological parameters in β-thalassaemia/HbE patients
The basic haematological parameters of the samples examined in this study are shown in Table I. Both normal controls and β-thalassaemia/HbE patients were matched in age. The splenectomised β-thalassaemia/HbE patients had strikingly abnormal values in platelet counts, nucleated RBCs, and Hb, which most probably reflected more severe haemolysis in these β-thalassaemia/HbE patients.
Table I. Haematological values in β-thalassaemia/Hb E patients and normal subjects.
β-thalassaemia/Hb E patients
RBC, red blood cell; MCV, mean corpuscular volume; MCH, mean corpuscular haemoglobin; MCHC, mean corpuscular haemoglobin concentration; WBC, white blood cell.
*Significantly different from normal subjects P < 0·0001.
No. of patients
28·02 ± 6·94
27·41 ± 7·27
27·14 ± 5·76
68·0 ± 17·0*
62·0 ± 13·0*
132·0 ± 12·0
0·214 ± 0·056*
0. 219 ± 0·034*
0·397 ± 0·037
RBC (× 1012/l)
3·60 ± 1·00*
3·30 ± 1·80*
4·60 ± 0·50
61·70 ± 7·30*
73·40 ± 10·30*
86·40 ± 5·40
19·10 ± 2·50*
20·90 ± 3·80*
28·60 ± 2·30
315·0 ± 23·0*
282·0 ± 25·0*
332·0 ± 10·0
WBC (× 109/l)
8·26 ± 6·39
11·63 ± 5·52*
5·89 ± 1·45
Platelet (× 109/l)
298·30 ± 290·51
651·89 ± 196·89*
230·89 ± 82·95
Nucleated RBC/100 WBC
20·00 ± 65·00
634·54 ± 436·93*
Percentage of annexin V+ MPs
Flow cytometric staining of peripheral blood showed that bivariate FSC/SSC efficiently distinguished among MPs, platelets and RBCs (Fig 1). The percentage of annexin V+ staining of MPs, platelet and RBC population were analysed by standard flow cytometry after staining the whole blood samples with FITC-conjugated annexin V. A high percentage of circulating MPs was found in blood samples from thalassaemia patients when compared with blood samples from normal subjects. The percentage of annexin V+ MPs (mean ± SD) in splenectomised β-thalassaemia/HbE patients was significantly higher than either non-splenectomised β-thalassaemia/HbE, P < 0·0001 or normal subjects, P < 0·0001 (11·66 ± 9·98%, 8·14 ± 9·11% and 6·97 ± 7·30% respectively Fig 2A with the median [IQR] of 10·86% [2·63], 7·87% [1·96], and 5·79% [1·57], respectively). Although there was a trend towards a higher percentage of annexin V+ MPs in non-splenectomised β-thalassaemia/HbE patients when compared with the normal group, this difference was not statistically significant (P = 0·08). The percentage of annexin V+ within the platelet population in splenectomised β-thalassaemia/HbE patients was also significantly higher when compared with non-splenectomised β-thalassaemia/HbE patients (P < 0·01) and normal subjects (P < 0·01) (12·01 ± 10·36%, 7·12 ± 6·83% and 4·29 ± 6·12%; median [IQR]: 12·66% [7·28], 6·97% [5·60], and 4·01% [4·10], respectively). In addition, splenectomised β-thalassaemia/HbE patients also had a significantly higher percentage of annexin V+ RBCs when compared with non-splenectomised β-thalassaemia/HbE patients (P < 0·05) and normal subjects (P < 0·01) (5·11 ± 3·85%, 0·82 ± 0·96% and 0·15 ± 0·10% with the median [IQR] of 4·96% [1·23], 0·85% [0·24], and 0·12% [0·09], respectively). Again no significant differences in the percentages of annexin V+ platelets and annexin V+ RBCs was found between non-splenectomised β-thalassaemia/HbE patients and those from the normal donors (P = 0·14 and P = 0·20) (Fig 2A).
Absolute number of annexin V+ MPs
The absolute numbers of annexin V+ MPs, platelets and RBCs, calculated as the number of annexin V-positive events present utilising the gating strategy outlined in the method section for MPs, platelets and RBCs with TruCountTM beads as an internal reference, were found to be significantly elevated in thalassaemia when compared with the normal group (Fig 2B). The number of total annexin V+ events in MPs, platelets and RBCs were markedly higher in the splenectomised β-thalassaemia/HbE patients when compared with non-splenectomised β-thalassaemia/HbE patients (P < 0·0001) and normal subjects (P < 0·0001) with the absolute values of 112 538·6 ± 62 499·1 events/μl (mean ± SD), 74 438·9 ± 127 461·3 events/μl, and 152 209·4 ± 99 214·8 events/μl with the median (IQR) of 86 104·6 events/μl (35 222·0), 35 816·6 events/μl (38 259·4), and 97 170·0 events/μl (52 617·5) for annexin V+ MPs, platelets and RBCs, respectively. In non-splenectomised β-thalassaemia/HbE patients, the absolute values of annexin V+ MPs and RBCs were also significantly higher (78 738·4 ± 42 843·1 events/μl, P < 0·01 and 29 173·8 ± 30 714·0 events/μl, P < 0·0001; median [IQR]: 66 301·0 events/μl [32 359·6] and 16 848·0 events/μl [14 436·0], respectively) when compared with the normal subjects (56 479·4 ± 17 328·9 events/μl, and 7084·8 ± 4808·9 events/μl with the median [IQR] of 53 915·6 events/μl [18 021·7], and 5040·0 events/μl [2769·0], respectively). There was no significant difference in the absolute values of annexin V+ platelets between non-splenectomised β-thalassaemia/HbE patients and normal subjects [8606·0 ±14 645·8 events/μl and 12 840·4 ± 11 120·0 events/μl, P = 0·536; median (IQR): 6668·0 events/μl (9573·1) and 7238·6 events/μl (5255·8)]. Interestingly, samples from some thalassaemia patients contained relatively higher values of annexin V+ MPs. A possible link between the levels of annexin V+ MPs and the annexin V+ in the platelet population was therefore examined and a positive correlation was observed with rs = 0·511, textitP < 0·001 (Fig 3A), and between the absolute number of annexin V+ MPs and annexin V+ RBCs (rs = 0·541, P < 0·001) (Fig 3B). Moreover, there was also a good correlation between platelet counts and platelet-derived MPs with rs = 0·580, P < 0·0001 (data not shown). For ITP patients, the number of annexin V+ MPs in peripheral blood was significantly lower than that in splenectomised β-thalassaemic patients (56 824·6 ± 18 857·6 events/μl vs. 99 795·6 ± 26 681·9 events/μl, P < 0·05; median [IQR] of 58 241·6 events/μl [25 757·2] vs. 105 153·0 events/μl [40 817·5]), but showed no difference when compared with normal controls (48 417·0 ± 14 425·8 events/μl (P = 0·423) with the median [IQR] of 50 357·7 events/μl [18 289·0]).
Cellular origin of MPs
To identify the cellular origin of MPs in the peripheral blood, a three-color flow cytometric assay was used to analyse aliquots of blood samples that were simultaneously stained with FITC-conjugated annexin V and PE- or PerCP or TC-conjugated mAbs against either platelet integrin αIIb (CD41a), circulating platelet antigen (CD36), activated platelet antigen (CD62P), RBC antigen glycophorin A, endothelial antigen endoglin (CD105/CD62E), a granulocyte marker (CD66e), a monocyte marker (CD14) or a leucocyte common antigen (CD45). The three-color flow cytometric assay demonstrated the presence of distinct populations of MPs in the samples of whole blood from both normal and thalassaemia patients. However, it is of interest to note that the numbers of annexin V+ MPs that expressed RBC antigen, CD62P, CD36 and CD41a were significantly higher when compared with other antigens (Fig 4A). The absolute number of MPs that expressed platelet origin (CD41a) and the procoagulant lipid PS (CD41a+/annexin V+) was significantly higher in splenectomised β-thalassaemia/HbE patients (mean absolute numbers: 22 825·7 ± 10 798·5 events/μl; median [IQR]: 22 888·5 events/μl [15 658·1]) and non-splenectomised β-thalassaemia/HbE (7836·5 ± 3119·5 events/μl with the median [IQR] of 7375·3 events/μl [3517·9]) when compared with normal subjects (6138·2 ± 3188·5 events/μl (Fig 4B) (P < 0·05, and P < 0·01, respectively), with the median (IQR) of 6124·4 events/μl (3372·1). A more pronounced difference was observed in platelet-derived MPs expressing activation marker (CD62P) (CD41a+/annexin V+/CD62P+). Thus, the activated platelet-derived MPs were markedly elevated in splenectomised β-thalassaemia/HbE patients (23 344·8 ± 11 312·3 events/μl; median [IQR]: 23 848·5 events/μl [15 060·8]) when compared with non-splenectomised β-thalassaemia/HbE patients (8670·5 ± 3450·2 events/μl; median [IQR]: 8311·4 events/μl [3880·8]) and normal subjects (7950·0 ± 2766·9 events/μl; median [IQR]: 7448·2 events/μl [2868·5]) (P < 0·05 and 0·01, respectively). Expression of the specific antigen (CD36) on platelet-derived MPs (CD41a+/CD36+) was also increased in the splenectomised β-thalassaemia/HbE patients (30 478·5 ±11 116·3 events/μl; median [IQR]: 30 342·4 events/μl [13 698·2]), and the non-splenectomised β-thalassaemia/HbE patients (11 712·3 ± 5103·6 events/μl; median [IQR]: 11 204·2 events/μl [7114·9]) when compared with the normal group (7812·9 ± 4348·4 events/μl; median [IQR]: 7880·4 events/μl [4154·0]). Annexin V+ MPs of RBC-derived, endothelial cell-derived, granulocyte-derived, leucocyte-derived, monocyte-derived and TF-derived antigens were also found in thalassaemia patients but at relatively lower levels when compared with CD41a+, CD62P+ and CD36+-derived MPs. Thus, the differences within the splenectomised β-thalassaemia/HbE, non-splenectomised β-thalassaemia/HbE, and the normal group for the former cell lineage-derived MPs did not reach statistical significance.
Platelet factor-3-like activity and the amount of MPs
To investigate whether the high levels of annexin V+ MPs detected in the thalassaemia blood samples promoted coagulation, we measured the activation of the coagulation system in vivo by assessing the platelet factor-3-like activity in whole blood. Whole blood from splenectomised β-thalassaemia/HbE patients had a very high availability of platelet factor-3 when compared with non-splenectomised β-thalassaemia/HbE patients (P < 0·05) and normal subjects (P < 0·01). OD405 =2·21 ± 0·76, 1·22 ± 0·49 and 1·13 ± 0·16, respectively (Fig 5) and the median (IQR) was 2·17 (0·77), 1·26 (0·61) and 1·10 (0·38), respectively. When a comparison was made between the number of annexin V+ MPs and the platelet factor-3-like activities, coefficients of determination from all blood samples were high (rs = 0·560; P < 0·001) (Fig 6).
An increasing number of studies have documented the presence of high levels of membrane-derived MPs in various vascular pathologies (Nieuwland et al, 1997; Miyamoto et al, 1998; Combes et al, 1999; Mallat et al, 1999; Vidal et al, 2001; Freyssinet, 2003; VanWijk et al, 2003). The high affinity binding of annexin V to MPs suggests that such a characteristic could contribute to defects in coagulation. However, to our knowledge, there has been no attempt to quantify the amount of MPs in thalassaemia blood. Although measurement of MPs using microplate assays with immobilised annexin V or cell-specific mAbs (Hugel et al, 1999; Mallat et al, 1999, 2000) or an ultra centrifugation method have been described (Nieuwland et al, 1997, 2000; Berckmans et al, 2001), these assays are not easy to perform because they involve laboratory manipulations at various steps of the assay and/or extensive ultracentrifugation, which could affect the interpretation of results because of substantial loss of MPs. Using a three-color flow cytometric technique, we were able to identify and quantify, for the first time, the amount of MPs in the peripheral blood of both normal individuals and thalassaemia patients. The direct immunostaining of whole blood and flow cytometric analysis, reported here, is a simple, rapid, reproducible, and practical clinical diagnostic technique. This technique also requires relatively small amounts of blood.
The data presented here clearly indicate that MPs are found in both normal and thalassaemia blood samples, but thalassaemia patients have significantly higher percentage and absolute number of MPs than normal subjects. The basal level of circulating MPs found in normal individuals probably reflects a normal physiological balance between cell activation and senescence or death of the cells. It could also reflect cells that have transiently escaped destruction by phagocytosis (Tanaka & Schroit, 1983; Schroit et al, 1985). Therefore, the increased amounts of MPs detected in the blood of thalassaemia patients, in which a high degree of cell apoptosis and senescence are noted, could be secondary to a saturation of the normal physiological process involved in the elimination of MPs. Such a defect may lead to increased levels of circulating-shed MPs. The higher amounts of MPs found in splenectomised patients than non-splenectomised patients indicate an important role of the spleen, which could play some part in the removal of these MPs. Cells that express PS on their surface (such as apoptotic and ageing cells) are known to be cleared by the splenic reticuloendothelial system. The spleen, liver and lung are the three major organs that have been shown to clear MPs in an animal model, in which the injection of mouse or human RBC led to the generation and clearance of MPs (Bocci et al, 1980). It is also known that splenectomy in β-thalassaemia uncovers a population of RBCs with particularly severe abnormalities that would otherwise have been selectively removed by the spleen (Eldor et al, 1991; Eldor & Rachmilewitz, 2002). This can result in increased numbers of damaged RBCs showing anisocytosis and poikilocytosis that can, in turn, generate a large amount of MPs in the circulation. It is important to note that the amount of platelet-derived MPs increases in β-thalassaemia patients postsplenectomy; whether this further increase in the levels of circulating MPs is of clinical significance remains to be established. This finding nonetheless suggests caution in using splenectomy as a therapeutic option as this surgical procedure might, in fact, exacerbate the detrimental effect of circulating MPs in such patients. In support of this view is the current discussion of the therapeutic value of splenectomy in patients with β-thalassaemia. Clearly, alternative strategies have to be devised to address this important issue.
Phenotypic analysis of the cellular origin of MPs showed that the largest population of MPs in the circulation was of platelet origin, and that these platelet-derived MPs predominantly expressed the aminophospholipid PS, and possessed the activation marker with proadhesive phenotypes (PS+/CD41+/CD62P+/CD36+). There are a number of studies showing that platelets are a major source of the generation of MPs in healthy individuals and in many patients who are at risk of thrombosis (Nieuwland et al, 2000; Berckmans et al, 2001; Freyssinet, 2003). This is at least partly because they are naturally relatively easy to be activated by several mechanisms involved in their specific function in haemostasis. Thus, abnormal thalassaemic RBCs with PS on the external membrane leaflet may enhance thrombin generation in vivo and trigger platelet activation (Eldor et al, 1999; Eldor & Rachmilewitz, 2002) followed by the generation of platelet-derived MPs. This suggests that a tendency towards the vesiculation of MPs from activated platelets may be a generalised characteristic of the stressed RBCs (Gail et al, 1986; Eldor & Rachmilewitz, 2002). The good correlation between PS+-RBCs and PS+-platelets found in this study helps support this view. The presence of PS on the outer surface of thalassaemic RBC and activated platelets may account for the elevated level of platelet factor 3-like activity in the whole blood of β-thalassaemia patients (Opartkiattikul et al, 1992). Platelet factor-3 plays a very important role in the activation of coagulation factors and is reasoned to facilitate the activation of platelets. This study has shown that a large proportion of PS+ platelet-derived MPs correlated well with the extent of platelet activation, as demonstrated by high levels of platelet factor-3 activity. MPs of activated platelet origin are of particular interest because of their high procoagulant potential. With regard to the chronic hypercoagulable state found in β-thalassaemia/HbE patients, there is evidence showing that, in thalassaemia, the platelets are in a state of chronic activation (Bocci et al, 1980; Eldor & Rachmilewitz, 2002) with an increased fraction of platelets expressing the activation markers CD62P and CD63 (Eldor et al, 1993; Eldor & Rachmilewitz, 2002). Moreover, our study supports the occurrence of increased circulating platelet aggregates in splenectomised β-thalassaemia/HbE patients, an observation compatible with in vivo platelet activation and the existence of a hypercoagulable state (Winichagoon et al, 1981). The fact that the phenotype of MPs was similar to that displayed by activated platelets, and the procoagulant PS and proadhesive character, suggests that MPs when in excess may represent a ‘circulating platelet compartment’ bearing the capacity to develop adhesive interactions with both neighboring or remote cells.
Although it is difficult to determine whether the increased amounts of MPs present in the circulation is sufficient to quantitatively cause thrombosis in thalassaemia, numerous reports of thromboembolic complications associated with certain haemolytic anaemias, in particular sickle cell anaemia, have been documented (Wagner et al, 1986; Eldor & Rachmilewitz, 2002). In thalassaemia, the presence of a higher than normal incidence of thromboembolic events and the existence of prothrombotic haemostatic anomalies are the consequence of changes in membrane phospholipid asymmetry of abnormal RBCs and activation of other blood cells, including platelets, monocytes, granulocytes and endothelial cells (Eldor & Rachmilewitz, 2002). Therefore, the presence of high levels of PS-bearing MPs originating from activated platelets might act as one of the confounding factors responsible for the dissemination of prothrombotic manifestations in thalassaemia.
This study was supported by the Thailand Research Fund-Senior Scholar and the Becton Dickinson Biosciences (Thailand). S.F. and K.P. are Senior Scholars of the Thailand Research Fund. We would like to thank Prof. Aftab A. Ansari for his valuable comments.