Circulating Leukocyte-Derived Microparticles Predict Subclinical Atherosclerosis Burden in Asymptomatic Subjects
Objective— To clarify circulating microparticles (MP) relationships with preclinical atherosclerosis.
Methods and Results— In 216 subjects without cardiovascular disease, we assessed: (1) annexin V-positive, platelet-derived, endothelium-derived and leukocyte-derived circulating MP by capture on annexin V, anti-GPIb, anti-CD105, and anti-CD11a antibody-coated wells, respectively; (2) Framingham risk, metabolic syndrome, and low-grade inflammation by risk factors measurement including hsCRP; and (3) subclinical atherosclerosis by ultrasound examination of carotid, abdominal aorta, and femoral arteries. Number of sites with plaque ranged from 0 to 3 and plaque burden was classified into 0 to 1 or 2 to 3 sites disease. Leukocyte-derived MP level was higher in the presence than in the absence of moderate to high Framingham risk (P<0.05), metabolic syndrome (P<0.01), high C-reactive protein (CRP) (P<0.05), or 2- to 3-sites disease (P<0.01), and correlated positively with number of metabolic syndrome components (P<0.001), tertiles of fibrinogen (P<0.001), and number of diseased sites (P<0.01). In multivariate analysis, 2- to 3-sites disease was independently associated with leukocyte-derived MP level (P<0.05), Framingham risk (P<0.001), and metabolic syndrome (P<0.01). None of the other MP types correlated with risk markers or atherosclerosis.
Conclusions— Leukocyte-derived MP, identified by affinity for CD11a, are increased in subjects with ultrasound evidence of subclinical atherosclerosis, unveiling new directions for atherosclerosis research.
Shed membrane microparticles (MP) in peripheral blood originate from diverse apoptotic or activated cells of endothelial, leukocyte, and platelet type and are involved in the pathogenesis of atherothrombotic disease.1,2,3 Indeed, MP isolated from blood of patients with clinically overt cardiovascular disease can initiate thrombosis and cause endothelial dysfunction in vitro and in experimental vascular models.4,5 Also, the number of circulating MP of endothelial and platelet origin is increased in patients with acute coronary syndromes6 or chronic kidney failure with vasculopathy.7 Nevertheless, little is still known about the potential participation of circulating MP to early stage of arterial disease before complication has occurred.1,3,8,9 To clarify this question we measured circulating MP level in symptom-free subjects without known history of cardiovascular disease and explored their potential relationships with cardiovascular risk factors and subclinical atherosclerosis assessed by ultrasound.10–12
The study population was constituted of consecutive asymptomatic subjects addressed between December 2004 and July 2005 to our Center of Cardiovascular Prevention by occupational health medicine doctors in the framework of a program of cardiovascular prevention described previously,10,13 aiming to determine their cardiovascular risk factor profile. A condition of selection was that they were free of any cardiovascular disease (stroke, coronary heart disease, heart failure, peripheral vascular disease, and diabetes), and that they had undergone measurement of circulating MP after written informed consent. Body mass index was calculated as the ratio of weight-to-height squared and waist circumference was measured with a tape horizontal around the abdomen at the level of the right iliac crest.14,15 Resting brachial blood pressure was measured by sphygmomanometer, and hypertension was defined as blood pressure ≥140/90 mm Hg and/or the presence of antihypertensive treatment.16 Fasting blood lipids and glucose were measured by enzymatic methods (after precipitation of low-density lipoprotein [LDL] and very-low-density lipoprotein [VLDL] for high-density lipoprotein [HDL] measurement), and hypercholesterolemia was defined by total cholesterol of ≥5.2 mmol/L, and/or the presence of lipid lowering drug treatment.16 Plasma high-sensitivity C-reactive protein (CRP) was measured by immunoassay17 and plasma fibrinogen was measured by the Claus method.18 Current smoking was defined by daily consumption of ≥1 cigarette for ≥3 months. Framingham risk of coronary heart disease (% at 10 years) was estimated by entering age, male sex, systolic pressure, total and HDL cholesterol, and presence or absence of smoking into the Framingham Model equations,19 and study subjects were classified as low risk (<10%) or moderate to high risk (≥10%). Metabolic syndrome was considered present when 3 of the following 5 criteria were present: (1) waist circumference >94 cm in men or >80 cm in women, in accordance with cut-offs for European subjects15; (2) triglycerides >1.71 mmol/L; (3) HDL cholesterol <1.03 mmol/L in men or <1.29 mmol/L in women; (4) blood pressure ≥130/85 mm Hg and/or on antihypertensive medication; and (5) fasting blood glucose ≥5.6 mmol/L.15 Inflammatory risk was graded according to three following levels of CRP: low CRP as <1 mg/L, intermediate CRP as ≥1 and <3 mg/L, and high CRP as ≥3 mg/L.20
MP were isolated from platelet-poor plasma and captured by immobilized specific anti-GpIb, anti-CD105 or anti-CD11a antibodies as previously described.21 The procoagulant phospholipid content was determined by a prothrombinase assay and the amount of MP (MP level) was expressed as nanomolar phosphatidylserine (PS) equivalent.
MP Sample Preparation
All samples were prepared and stored as follows. Five milliliters of 12-hour fasting venous blood were withdrawn on citrated tube. Within the 2 hours after withdrawal, samples were centrifuged at 1500g for 15 minutes at room temperature, and the supernatant was centrifuged again at 13 000g for 2 minutes to avoid platelet contamination. The so-obtained platelet-poor plasmas were distributed into 4 aliquots of 500 μL each, frozen, and stored at −80°C for a maximum of 2 months. After thawing, thrombin and factor Xa inhibitors, respectively, D-phenylalanyl-prolyl-arginyl chloromethyl ketone, and 1,5-dansyl-glutamyl-glycyl-arginyl chloromethyl ketone (both from Calbiochem, San Diego, Calif) were added to plasma samples at a final concentration of 50 μmol/L each and CaCl2 at a final concentration of 50 mmol/L.
Circulating Procoagulant MP Quantitation
After capture of MP on annexin V-coated wells (for 30 minutes at 37°C), taking advantage of the strong affinity of annexin V for aminophospholipids, chiefly PS, present in MP at the calcium concentration used, 4 washing steps were performed with Tris buffer containing 1 mmol/L CaCl2 and 0.05% Tween 20, each for 5 minutes at 20°C, and the last 1 without Tween. The PS content of MP, directly responsible for their procoagulant activity, was then measured in a prothrombinase assay in which the amount of generated thrombin is proportional to PS concentration. MP were incubated with factor Xa (50 pmol/L), factor Va (360 pmol/L), prothrombin (1.3 μmol/L), and 2.3 mmol/L CaCl2 for 15 minutes at 37°C. The reaction was stopped by an excess of EDTA, and linear absorbance changes were recorded at 405 nm after addition of Chromozym TH (Roche Diagnostics, Mannheim, Germany; 380 μmol/L) and converted to thrombin concentration by using a reference curve constructed with liposomes of known PS concentrations.21 Intra-assay and inter-assay coefficients of variation of MP measurement were ≈10% and ≈20% respectively, over a rather wide concentration range, eg, between 1 and ≈10-fold the normal.22
Platelet-Derived MP Quantitation
After specific capture of platelet-derived MP on anti-GPIb antibody-coated wells, quantitation was achieved after several washing steps using the prothrombinase assay, as described.
Endothelium-Derived MP Quantitation
After specific capture of endothelium-derived MP on anti CD105 antibody-coated wells, quantitation was achieved after several washing steps using the prothrombinase assay, as described.
Leukocyte-Derived MP Quantitation
After specific capture of leukocyte-derived MP on anti-CD11a antibody-coated wells, quantitation was achieved after several washing steps using a prothrombinase assay, as described.
Likewise with enzyme-linked immunosorbent assay (ELISA) methods, because incubation times and affinities for the various insolubilized ligands are different, no direct comparison can be made between amounts of MP captured onto annexin V or antibodies to particular phenotypes, without affecting the significance of the relative variations within each of the corresponding categories.
The presence of plaque at 3 different arterial sites (extracranial carotid arteries, abdominal aorta, and femoral arteries) was detected according to a procedure described and validated previously.10,13,18,23 Plaque was detected by high-resolution ultrasound (ATL 5000; Philips), using a probe with an operating frequency of 5 to 12 MHz for carotid and femoral investigations and of 5 MHz for abdominal aorta, and defined as a focalized encroachment into lumen by >1.5 mm. Each of the 3 sites investigated was characterized either by the absence of any plaque, or by the presence of at least 1 plaque regardless the number and location of such plaques. This dichotomous characterization of absence or presence of plaque allowed to obtain 94% to 100% agreement between repeated examinations.10 Among the 3 sites investigated, the number of sites with presence of plaque was coded as grade 0, 1, 2, or 3.10 A categorical estimation of 3 sites plaque burden was obtained by regrouping grades 0 and 1 (0 to 1 site disease) and grades 2 and 3 (2 to 3 sites disease).10,13
Data are presented as means (SD) for continuous variables and as number of subjects with percentage for qualitative variables. Log-transformed MP levels were compared between groups by ANOVA and additional nonparametric tests (Wilcoxon) were performed without log-transformation of MP. Associations between log-transformed MP and continuous parameters were analyzed by linear regression by least square method. Proportions of subjects with 2 to 3 sites disease were compared between subgroups by χ2 test. Independent associations of MP with atherosclerosis were tested by logistic regression entering 2- to 3-sites disease as dependent variable, and log-transformed MP level and covariates as independent variables. Statistical significance was set at P<0.05.
Two hundred sixteen subjects were eligible for the study. There was a high prevalence of hypertension (40%), hypercholesterolemia (39%), and smoking (23%). The presence of moderate to high Framingham risk was found in 44% of patients, that of metabolic syndrome in 37%, and high level of CRP in 18%. Approximately half of the subjects had 2- to 3-sites disease. Subjects characteristics are detailed in Tables 1 and 2⇓.
MP and Circulating Blood Cells
Leukocyte-derived MP level was positively associated with leukocyte cell count (r=0.52, P<0.001). Annexin V-positive and platelet-derived MP levels were positively associated with platelet count (r=0.23, r=0.24 respectively, P<0.001). None of the MP categories were associated with hematocrit.
MP and Cardiovascular Risk
No significant difference existed in annexin V-positive, platelet-derived and endothelium-derived MP levels between subjects with moderate to high Framingham risk and those with low Framingham risk, or between subjects with metabolic syndrome and those without, or between subjects with high CRP and those with low or intermediate CRP (Table 2). In contrast, leukocyte-derived MP level was higher in subjects with moderate to high Framingham risk than in low Framingham risk subjects (P<0.05) and in subjects with metabolic syndrome than in those without (P=0.001) (Table 2). Also, leukocyte-derived MP level differed between the 3 grades of CRP (P<0.05), with higher value in subjects with high CRP than in those with low CRP (P<0.01) (Table 2). It is noteworthy that similar levels of statistical significance were obtained by using Wilcoxon test without log-transformation of MP (respectively, P<0.05, P<0.001, P<0.05). Last, leukocyte-derived MP level was positively related with increasing number of components of metabolic syndrome (r=0.23, P<0.001; Figure 1) and tertiles of plasma fibrinogen (r=0.31, P<0.001; Figure 1).
MP and Subclinical Atherosclerosis
No significant difference in annexin V-positive, platelet-derived, and endothelium-derived MP levels existed between subjects with 2- to 3-sites disease and those with 0- to 1-site disease (Table 2). In contrast, leukocyte-derived MP level was higher in subjects with 2- to 3-sites disease than in those with 0- to 1-site disease (P<0.01; Table 2), with a similar difference obtained by using Wilcoxon test without log-transformation of MP (P<0.05). A positive relationship existed between leukocyte-derived MP level and the number of sites with presence of plaque (r=0.20, P<0.01; Figure 1). When study subjects were divided into 4 subsets based on leukocyte-derived MP level (low, <, or high, ≥, median value, 3.3 nmol/L PS) and Framingham risk score (low, <, or high, ≥, median value, 8.6%), the prevalence of 2- to 3-sites disease was greatest (70%) in individuals with high Framingham risk score and high leukocyte-derived MP level, and lowest (27%) in those with low Framingham risk score and low leukocyte-derived MP level (P<0.001; Figure 2). Multivariate regression models, adjusted for leukocyte cell count and coexisting antihypertensive and lipid-lowering therapy, showed that the presence of 2- to 3-sites disease was independently associated with leukocyte-derived MP level (P<0.05), Framingham risk score (P<0.001), and presence of metabolic syndrome (P<0.01), but not CRP (model 1) or fibrinogen (model 2) (Table 3).
The present investigation clearly evidenced a relationship between the levels of circulating leukocyte-derived MP and the degree of cardiovascular risk markers and atherosclerotic plaque burden. It has to be emphasized that MP values of our low-risk low-inflammation subgroups without atherosclerosis were not different from those observed in healthy individuals in our routine practice (unpublished data, Table 2).
The involvement of leukocyte-derived MP in atherosclerosis has been previously reported in symptomatic patients with clinically overt arterial disease, whose atherosclerotic plaques contained great amounts of leukocyte-derived MP,24 but their participation to the preclinical early stage of atherosclerotic disease is suggested for the first time by the present study. Ultrasound interrogation of 3 arterial sites, carotid arteries, abdominal aorta, and femoral arteries, in a cohort of asymptomatic subjects without cardiovascular disease, allowed us to find that leukocyte-derived MP level, measured by CD11a antibody in peripheral blood, was higher in subjects carriers of atherosclerotic plaque in 2 or 3 sites than in those without plaque at any site or with plaque in only 1 site. Moreover, circulating leukocyte-derived MP level increased significantly, although weakly, gradually in function of the presence of plaque in 0, 1, 2 or 3 sites. The positive association between leukocyte-derived MP level and presence of high degree of plaque burden (2 to 3 sites disease) was independent of coexisting cardiovascular risk factors, circulating leukocyte cell count and coexisting anti-hypertensive or lipid-lowering drugs. These findings suggest some etiologic link between leukocyte-derived MP and the burden of subclinical atherosclerosis within the arterial tree.25 They also allow to consider leukocyte-derived MP as potential valuable predictors for subclinical atherosclerosis burden, with a predictive value, estimated by odds ratio of having 2- to 3-sites disease in multivariate regression analysis, similar to that of metabolic syndrome and CRP but half lesser than that of Framingham risk score. Anyway, as illustrated by Figure 2, leukocyte-derived MP level adds a useful predictive value to that of Framingham risk score as regards the risk of diffuse atherosclerosis. The link of leukocyte-derived MP with plaque burden being independent of conventional cardiovascular risk factors, specific deleterious effects of leukocyte-derived MP on the arterial wall are likely to occur. One possible effect, supported by a study of circulating leukocyte-derived MP in normal volunteers,9 may be to stimulate the release of inflammatory cytokines and the expression of functional tissue factor in endothelial cells, that may potentially contribute to endothelial dysfunction. Another effect, reported in mouse and concerning MP derived from T lymphocytes, may be to induce dysregulation of endothelial function of conductance arteries by affecting nitric oxide synthase.26
Circulating leukocyte-derived MP level was also influenced by various cardiovascular risk markers such as metabolic syndrome, low-grade inflammation, or Framingham risk score, independently of its relationship with atherosclerotic plaque burden. Although the associations between leukocyte-derived MP level and atherosclerotic risk factors were somewhat weak, they remained statistically significant whether data were analyzed by using log transformation or parametric testing. In subjects with metabolic syndrome, leukocyte-derived MP level was higher than in those free of such syndrome and in the overall study population, leukocyte-derived MP level increased gradually in parallel with the number of components of metabolic syndrome. These findings agree with previous studies showing that type 2 diabetic patients present an increased number of circulating MP of leukocyte origin.27,28 Leukocyte-derived MP level was also increased in the presence of low-grade inflammation, in agreement with a recent work showing that circulating MP derived from leukocyte of healthy subjects are rapidly up regulated by inflammatory stimulations.9 Last, leukocyte-derived MP level was increased in the presence of moderate to high Framingham risk score, showing that leukocyte-derived MP may reflect the influence of integrated and cumulative effects of traditional risk factors.
The lack of association between circulating procoagulant, platelet-derived, and endothelium-derived MP levels and plaque burden, Framingham risk, metabolic syndrome, or low-grade inflammation constitutes a last series of findings that merits discussion. They are discrepant with the results of previous studies in patients with clinically advanced arterial disease showing that myocardial infarction, coronary artery disease, or end-stage renal failure were associated with increased number of endothelium and/or platelet-derived MP.6,7 They also contradict a previous observation that the number of endothelium-derived MP was positively related with most traditional cardiovascular risk factors in patients with coronary artery disease.29 Such discrepancies may be explained by the fact that platelet-derived MP exert their major effect on thrombogenesis, which is absent or minimal at the preclinical stage of arterial disease before atherothrombotic complication. This explanation applies also to circulating procoagulant MP that are the preponderant reflection of platelet-derived MP. Moreover, the lack of association of endothelium-derived MP level with cardiovascular risk or preclinical atherosclerosis may be caused by the weak endothelial injury and apoptosis in our asymptomatic subjects at early stage of arterial disease. However, it cannot be excluded that the discrepancies between our present data and previous studies showing increased endothelial-derived MP level in various clinical conditions6,7,29 result from methodological differences between the cited studies and ours. In one case, endothelial markers used for immunocapture assay were different from those used in our study;6 in others, MPs were detected by fluorescence activated cell sorter (FACS) analysis instead of immunocapture assay.7,29 Other studies in asymptomatic subjects have shown that endothelial-derived MPs were increased, either acutely after a single ingestion of fatty meal,30 or chronically in presence of metabolic syndrome.31 This discrepancy with our findings could again arise from different assay system, because the mentioned studies used FACS analysis instead of solid-phase capture assays.
The lack of increase of annexin V-positive MP with atherosclerotic risk, contrasting with the increasing leukocyte-derived MP level with the degree of cardiovascular risk, may be accounted for by the fact that leukocyte-derived MP represent only a fraction of total PS-expressing MP, as measured by annexin V binding. If leukocyte-derived MP is the sole subpopulation of MP that changed with risk factors, this may not be sufficient to significantly affect the total annexin V-positive content.
MPs were detected by using microplate affinity capture assays, which do not provide information about MP size and detect only MP which expose functional PS on their surface, depending on activation versus apoptosis. As a result, it cannot be ruled out that some annexin V-negative MP contribute to increased atherosclerotic risk. Further studies using different approaches to detect MP, including FACS analysis, are required to address this possibility. Also, we cannot exclude that a lack of sensitivity of MP markers may be a limitation for analyzing our asymptomatic subjects. This may contribute to the lack of relationship between platelet- and endothelial-derived MP and atherosclerotic risk markers.
In conclusion, circulating leukocyte-derived MP are independently associated with subclinical atherosclerosis burden, and this may unveil an important new direction for atherosclerosis research.
We thank Dr Yann Bihorel for his invaluable assistance in clinical data collection.
Sources of Funding
This work was supported by a grant from the Agence Nationale de la Recherche (Projet MIPRA-Met; ANR-05-PCOD-24-01) and from Sanofi-Aventis France (Institut d’Athérothrombose). A.T. was supported by a Contrat d’Interface Inserm.
G.C. and A.S. contributed equally to this study.
Original received June 26, 2006; final version accepted September 25, 2006.
Hugel B, Martinez MC, Kunzelmann C, Freyssinet JM. Membrane microparticles: two sides of the coin. Physiology. 2005; 20: 22–27.
VanWijk M, VanBavel E, Sturk A, Nieuwland R. MP in cardiovascular diseases. Cardiovasc Res. 2003; 59: 277–287.
Martinez MC, Tesse A, Zobairi F, Andriantsitohaina R. Shed membrane microparticles from circulating and vascular cells in regulating vascular function. Am J Physiol Heart Circ Physiol. 2004; 288: H1004–1009.
Boulanger C, Scoazec A, Brahimian T, Henry P, Mathieu E, Tedgui A, Mallat Z. Circulating microparticles from patients with acute myocardial infarction cause endothelial dysfunction. Circulation. 2001; 104: 2649–2452.
Mallat Z, Benamer H, Hugel B, Benessiano J, Steg PG, Freyssinet JM, Tedgui A. Elevated levels of shed membrane microparticles with procoagulant potential in the peripheral circulating blood of patients with acute coronary syndromes. Circulation. 2000; 101: 841–843.
Amabile N, Guérin AP, Leroyer A, Mallat Z, Nguyen C, Boddaert J, London GM, Tedgui A, Boulanger CM. Circulating endothelial microparticles are associated with vascular dysfunction in patients with end-stage renal failure. J Am Soc Nephrol. 2005; 16: 3381–3388.
Mesri M, Altieri DC. Leukocyte MP stimulate endothelial cell cytokine release and tissue factor induction in a JNK1 signaling pathway. J Biol Chem. 1999; 274: 23111–23118.
Simon A, Giral P, Levenson J. Extracoronary atherosclerotic plaque at multiple sites and total coronary calcification deposit in asymptomatic men. Association with coronary risk profile. Circulation. 1995; 92: 1114–1121.
Simon A, Megnien JL, Levenson J. Coronary risk estimation and treatment of hypercholesterolemia. Circulation. 1997; 96: 2449–2452.
Devynck MA, Simon A, Pernollet MG, Gariepy J, Rendu F, Levenson J. Plasma cyclic GMP and large artery remodeling in asymptomatic men. Hypertension. 2004; 44: 919–923.
Grundy SM, Cleeman JI, Daniels SR, Donato KA, Eckel RH, Franklin BA, Gordon DJ, Krauss RM, Savage PJ, Smith Jr SC, Spertus JA, Costa F. Diagnosis and management of the metabolic syndrome. An Am Heart Association/National Heart, Lung, and Blood Institute scientific statement. Circulation. 2005; 112: 2735–2752.
National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation. 2002; 106: 3143–3421.
Levenson J, Giral P, Megnien JL, Gariepy J, Plainfosse MC, Simon A. Fibrinogen and its relations to subclinical extracoronary and coronary atherosclerosis in hypercholesterolemic men. Arterioscler Thromb Vasc Biol. 1997; 17: 45–50.
Pearson TA, Mensah GA, Alexander RW, Anderson JL, Cannon RO 3rd, Criqui M, Fadl YY, Fortmann SP, Hong Y, Myers GL, Rifai N, Smith SC, Jr., Taubert K, Tracy RP, Vinicor F. Markers of Inflammation and Cardiovascular Disease. Application to Clinical and Public Health Practice: A Statement for Healthcare Professionals From the Centers for Disease Control and Prevention and the Am Heart Association. Circulation. 2003; 107: 499–511.
Weinstabl H, Vormittag R., Sailer T., Koder S., Freyssinet J.M., Pabinger I. Adaptation of a kinetic assay for the determination of blood circulating microparticles. Hämostaseologie. 2006; 26: 288(abstr).
Chironi G, Gariepy J, Denarie N, Balice M, Megnien JL, Levenson J, Simon A. Influence of hypertension on early carotid artery remodeling. Arterioscler Thromb Vasc Biol. 2003; 23: 1460–1464.
Mallat Z, Hugel B, Ohan J, Lesèche G, Freyssinet JM, Tedgui A. Shed membrane microparticles with procoagulant potential in human atherosclerotic plaques. A role for apoptosis in plaque thrombogenicity. Circulation. 1999; 99: 348–353.
Stamler J. Research related to risk factors. Circulation. 1979; 60: 1575–1587.
Martin S, Tesse A, Hugel B, Martinez MC, Morel O, Freyssinet JM, Andriantsihohaina R. Shed membrane particles from T lymphocytes impair endothelial function and regulate endothelial protein expression. Circulation. 2004; 109: 1653–1659.
Sabatier F, Darmon P, Hugel B, Combes V, Sanmarco M, Velut JG, Arnoux D, Charpiot P, Freyssinet JM, Oliver C, Sampol J, Dignat-George F. Type 1 and type 2 diabetic patients display different patterns of cellular MP. Diabetes. 2002; 5: 2840–2845.
Diamant M, Nieuwland R, Pablo RF, Sturk A, Smit JWA, Radder JK. Elevated numbers of tissue-factor exposing microparticles correlate with components of the metabolic syndrome in uncomplicated type 2 diabetes mellitus. Circulation. 2002; 106: 2442–2447.
Werner N, Wassmann S, Ahlers P, Kosiol S, Nickening G. Circulating CD31+/annexin V+ apoptotic MP correlate with coronary endothelial function in patients with coronary artery disease. Arterioscler Thromb Vasc Biol. 2006; 26: 112–116.
Ferreira AC, Peter AA, Mendez AJ, Jimenez JJ, Mauro LM, Chirinos JA, Ghany R, Virani S, Garcia S, Horstman LL, Purow J, Jy W, Ahn YS, de Marchena E. Postprandial hypertriglyceridemia increases circulating levels of endothelial cell microparticles. Circulation. 2004; 110: 3599–3603.