Microparticle increase in severe obesity: Not related to metabolic syndrome and unchanged after massive weight loss

Authors


  • Funding agencies: This study was supported by AP-HP DRCD (FAP04014) and AP-HP, hôpital Louis Mourier.

  • Disclosure: The authors declared no conflict of interest.

Abstract

Objective

To clarify the relationships between circulating microparticles (MPs), leukocyte–platelet aggregates (LPAs), obesity, and metabolic abnormalities and evaluate the effect of losing weight on these parameters.

Design and Methods

In 151 severely obese women and 60 lean controls, total MPs (annexin-V positive), platelet (annexin V/CD41+) and endothelial (CD31+/CD41−) MPs, and LPAs using flow cytometry, and the presence of metabolic syndrome (MS) were assessed. The effect of weight loss was studied in 32 subjects after a 1 year follow-up.

Results

Total microparticle count, platelet MPs (PMPs) and endothelial MPs (EMPs), and neutrophil–platelet aggregates were significantly increased in obese subjects versus lean controls, independently of the MS. Within obese subjects, there was no significant difference between those having and those not having MS. MPs and LPA counts did not vary significantly in subjects who lost 25% of their excess weight.

Conclusions

PMPs and EMPs, and LPAs are associated with obesity independently of metabolic abnormalities, but do not significantly change after massive weight loss. Further studies are needed to evaluate the prognostic significance of these observations, as beneficial effects of MPs are currently reported in addition to their initially described deleterious effects.

Introduction

Obesity has reached epidemic proportions in the world. In the United States, according to three surveys, the crude prevalence estimates of obesity are close to 29% [1]. Obesity is a major cardiovascular risk factor. This increased risk is in part linked to the increased prevalence of the metabolic syndrome (MS), a cluster of well-known cardiovascular risk factors (diabetes, hypertension, dyslipidemia). Interestingly, a unique subset of obese individuals has been described in the literature that appears to be protected from development of metabolic disturbances associated with obesity. These individuals termed Metabolically Healthy but Obese (MHO) in contrast with Metabolically Abnormal Obese (MAO) exhibit a normal metabolic profile [2]. Although a matter of controversy, the recent report of the American Heart Association [3] clearly indicates that obesity is associated with an increased risk of cardiovascular mortality and morbidity and that this risk is independent of classical cardiovascular risk factors. Thus, MHO subjects also have an increased cardiovascular risk. However, the mechanisms underlying this increased risk are not fully defined. A marker of this risk among obese patients would be of great interest.

Activated or apoptotic blood cells release in the circulation cellular microparticles (MPs) mainly originating from platelets but also from other cell types including endothelial cells, leukocytes, and vascular smooth muscle cells [4]. MPs expose proteins that are specific to the cell that they stem from; interestingly, they have procoagulant properties as they express anionic phospholipids and occasionally tissue factor. Most of the experimental evidence indicates that MPs influence diverse biological functions. They have a potent pro-inflammatory effect and their role in thrombus initiation and progression has been demonstrated [5]. Several studies have also shown the association between plasma concentrations of specific MPs and major types of endothelial dysfunction, namely inappropriate or maladaptative vascular tone, leukocyte recruitment, and thrombosis [6], processes which are involved in the pathophysiology of cardiovascular diseases [6, 7]. Increased levels of circulating MPs have been reported in association with acute coronary syndromes, diabetes, and within the atherosclerotic plaque. Circulating MPs are increased in most cardiovascular diseases and in patients with cardiovascular risk factors [4]. The conjugation of leukocytes with platelets leading to leukocyte–platelet aggregates (LPAs) is promoted by different conditions associated with inflammation and endothelial dysfunction. A clear association of monocyte–platelet–aggregates (MPA) with the onset of atherothrombosis has been reported [8]. Adiponectin is produced by adipose tissue; it has antithrombotic properties and inhibits platelet aggregation and is decreased in obese subjects [9]. In a recent study, increased adiponectin levels were shown to be independently associated with a reduced risk of acute coronary syndrome [10]. Adiponectin could be involved in MP production, but relationship between MPs and adiponectin has never been studied. The major independent roles shown for these markers in cardiovascular diseases strongly suggest the potential interest of studying their inter-relationship and link with obesity.

In obese and overweight subjects, two pilot studies, involving a small number of subjects (50 cases and 50 controls), reported an increased number of circulating MPs [11, 12]. In the latter study, the cellular origin of MPs was not identified. Our work hypothesis is that MPs and LPAs are increased in some obese patients and that this increase might help identifying a subgroup of subjects with markedly increased cardiovascular risk, notably in young women with severe or massive obesity.

The main objectives of this study were (i) to compare MPs and LPAs between severely obese subjects and subjects having a normal weight; and (ii) to analyze the relationship between these markers of cell activation and the MS. The secondary objective was to evaluate the effect of weight loss on MPs and LPAs concentrations. The distribution of adipose tissue and the prevalence of MS, classically lower in women [13], are different according to gender. We suspect that the increase in MPs also relies on different mechanisms in men and women. Considering the large predominance of women in our patients, we decided to perform this study in women only.

Methods

Study population

One hundred and fifty-one obese women, selected among those consecutively enrolled in COLOMBES cohort, a cohort dedicated to the study and follow-up of obese subjects, were evaluated between 2005 and 2009, up to once a year. Inclusion criteria of COLOMBES cohort (registration number in clinical trials registry NCT00632671) were: age ≥ 18 and < 60 years, body mass index (BMI) equal to or higher than 35 kg/m2, stable weight (defined as variations in weight of less than 10% over a period of 3 months before inclusion) and possible follow-up. Non-inclusion criteria were: pregnancy or breastfeeding, previous bariatric surgery, positive HIV serology, and anticancer treatment on course or oral corticoid treatment. The cohort was approved by the Institutional Committee of Ethical Practice (CCPPRB) of the Hôpital Henri Mondor in Créteil (France) and each subject gave informed written consent. In our study, subjects had to be free of any known cardiovascular disease (stroke, coronary heart disease, and peripheral vascular disease), thrombo-embolic disease, hypertension, dyslipidemia, diabetes, or intercurrent disease. They were not treated for hypertension, diabetes, or dyslipidemia, and did not receive anti-platelet agents. Among obese patients, 34 underwent bariatric surgery (18 gastric bypass and 16 gastric banding) during the study; a new clinical and biological evaluation was performed in 84 subjects one year after inclusion or surgery. A successful loss of weight was defined as a reduction of 25% of excess weight when compared to the inclusion evaluation. Excess weight was defined as the difference between the ideal weight and the current weight. Ideal weight (kg) was obtained using Lorentz formula (height (cm) − 100 − [(height (cm) – 150)/2.5]).

Sixty control women, aged between 18 and 60, were volunteers recruited among the medical and paramedical staff of our institution. They also gave informed written consent. Their BMI had to be greater than or equal to 18.5 and lower than 25 kg/m2. Other inclusion criteria were the same as for the obese group.

Anthropometry and blood pressure determination

Height and weight were measured with standard methods for patients and controls. Two brachial cuff blood pressure recordings were obtained at 5-min intervals and the mean value was entered. An obese-cuff was used when necessary. Smoking was defined as cigarette use within the previous 30 days. MS was defined according to the International Diabetes Federation (IDF) criteria [14]: waist circumference > 80 cm plus at least two of the following abnormalities: fasting glycaemia ≥ 5.5 mmol/l or diabetes therapy, systolic/diastolic blood pressure ≥ 130/85 mmHg or antihypertensive medication, triglycerides ≥ 1.7 mmol/l or lipid-lowering therapy, high density lipopretein (HDL)-cholesterol < 1.30 mmol/l or lipid-lowering therapy.

MP preparation and quantification

Blood samples were drawn after an overnight fast into 0.129 mol/l tri-sodium citrate tubes (Vacutainer, Becton-Dickinson, France) from a peripheral vein using a 21-gauge needle, and prepared within 2 h of collection according to the International Society of Thrombosis and Haemostasis (ISTH) recommendations for MP analysis, and previously published protocols [15]. Briefly, samples were centrifuged 15 min at 1,500g at 20°C, plasma was then harvested and centrifuged 2 min at 13,000g to remove all residual platelets. MP-containing plasma samples were stored frozen at −80°C and thawed at 37°C before analysis. Quantification or MPs was processed by flow cytometry (CXP FC500, Beckmann Coulter) according to their size and fluorescence. Three size of beads (3, 0.9, and 0.5 μm) (Megamix beads, Biocytex, Marseille, France) were used as reference for size determination. MPs were counted within region below 0.9 μm in size. For MPs labeling, 30 μl of plasma were incubated with specific monoclonal antibodies (annexin V–fluorescein isothiocyanate (FITC), anti-CD31-FITC (Beckman Coulter, Villepinte, France), anti-CD41–phycoerythrin (PE) (Immunotech, Marseille, France)), or corresponding isotype controls (isotypic IgG1-PE and IgG1-FITC (Immunotech, Marseille, France)). After 30 min of incubation, samples were diluted in phosphate buffered saline or binding buffer and internal standard (Flowcount beads, Beckmann Coulter, Villepinte, France) was added to express counts of MP as absolute numbers (Figure 1). Within total MP (defined as annexin-V positive), platelet MPs (PMPs) were defined as CD41+. Endothelial MPs (EMPs) were defined as CD31+/CD41− [16, 17]. Results were expressed as number of MP/μl of plasma.

Figure 1.

Representative graphs of cytofluorometry analysis of circulating MPs and PMP. (A) Three size of Megamix beads (3, 0.9, and 0.5 μm) were used as reference for size determination; they are first recognized on the basis of their side scatter (SS) and FL1 fluorescence properties. (B) The microparticle (MP) region was constructed on a SS log × FS log; MPs are defined as events with size below 1 μm using the 0.9 μm bead autogate. (C) Total MPs are defined as annexin V–FITC positive labeled events; platelet microparticles (PMP) are counted using dual fluorescence analysis using annexin V–FITC (FL1) and anti-CD41–PE (FL2).

Procoagulant MPs were quantified according to Hugel et al. [19], using the Zymuphen MP Activity® kit, manufactured by Hyphen (BioMed, Neuville-sur-Oise, France).

Neutrophil–platelet and monocyte–platelet aggregates preparation and quantification

Blood samples were drawn into 0.129 mol/l CTAD tubes (Diatubes-H, Franconville, France) and immediately brought to the laboratory. Blood cells were fixed in 5% paraformaldehyde (100 μl paraformaldehyde in 900 μl whole blood) and processed within 30 min; in parallel, blood cell count was made to quantify platelets, Neutrophils, and monocytes (Sysmex analyser). Neutrophil–platelet aggregates (NPA) were labeled with anti-CD41-PE and anti-CD16-PE/Cy5 (PC5), or an anti-IgG-PE isotypic control; MPA were labeled with anti-CD41-PE and anti-CD14-PC5 or an anti-IgG-PE isotypic control.

Other laboratory assays

Measurement of fasting glucose, total and HDL-cholesterol, triglycerides, insulin, creatinine, microalbuminemia, fibrinogen, and high-sensitivity C-reactive protein (hs-CRP) were performed using standard methods. Homeostasis model assessment (HOMA-R) was calculated according to the formula of Matthews et al. [20]. Serum adiponectin was measured using the radioimmunoassay (RIA) kit purchased from Linco Research (St Charles MO-6304, USA).

Statistical analysis

The statistical analysis was performed by ALTIZEM (Nanterre, France). Demographic data and medical characteristics were described at baseline. Subject data at inclusion are presented separated by case/control status, as well as by the presence (MS+) or absence (MS−) of MS. Qualitative values are reported as percentages and quantitative values as median and interquartiles. All cytometric parameters had log-normal distributions and were subsequently transformed before analysis. Comparison of quantitative variables between obese and non-obese subjects was performed using Student test or Mann and Whitney test, and using conditional logistic regression analysis after accounting for age, platelet, monocyte and neutrophil counts, adiponectin and CRP levels, smoking, blood pressure, presence of MS, HOMA index, hormonal treatment, or estro-progestative contraception. Associations between levels of MPs and obesity, as well as the subsidiary variables, were in all cases assessed by means of bivariate or multivariate linear regressions.

With 151 obese patients and 60 non-obese control women, our study is powered to have a 90% chance of detecting a difference of 1,250 MP/μL for total or PMPs (standard deviation: 2,500), and 6 EMPs/μl (standard deviation of 12). Thus, the effect size is 0.5, which is considered as moderate. The relationship between levels of MPs or aggregates and the presence (MS+) or absence (MS−) of the MS was further explored within the obese subpopulation via a multivariate regression, adjusting for the effect of several factors (waist-hip ratio, creatinine, microalbuminuria, hs-CRP, fibrinogen, and HOMA index). For each cytometric parameter, the influence of subjects' characteristics was studied thanks to an ANCOVA. Comparison of MPs and LPAs between obese patients who lost weight and controls was performed with ANCOVA after adjustment for previously described factors.

Results

Study population

Table 1 shows basic clinical and metabolic characteristics of the studied subjects. Compared with lean women, obese women had greater anthropometric measurements (BMI, waist circumference, waist to hip ratio [WHR]) and higher blood pressure, higher blood glucose, higher blood insulin, and lower HDL-cholesterol. CRP levels were higher and adiponectin levels were lower in obese versus lean women. The MS was found in 41% of obese women and in no lean women.

Table 1. Clinical and metabolic characteristics of studied women
 Non-obese, (n = 60)All obese, (n = 151)Obese without MS, MS− (n = 89)Obese with MS, MS+ (n = 62)
  1. Data are presented as number (%) or median and range between 25th and 75th percentiles. MS: metabolic syndrome; BMI: body mass index.
Age (years)39 (30; 47)33 (28; 44)34 (29; 45)32 (27; 40)
BMI (kg/m2)22 (20; 23)41.82 (39; 48)41 (39; 46)43 (41; 49)
Waist circumference (cm)72 (70; 76)114 (109; 125)113 (109; 124)116 (110; 128)
Waist to hip ratio (WHR)0.77 (0.74; 0.80)0.86 (0.82; 0.91)0.85 (0.82; 0.91)0.86 (0.82; 0.91)
Current smoking, n (%)13 (22.0)33 (22.0)19 (21.6)14 (22.6)
Blood pressure (mmHg)
Systolic119 (110; 125)128 (120; 130)120 (115; 130)130 (130; 135)
Diastolic80 (65; 80)70 (70; 80)70 (70; 80)76 (70; 80)
Blood lipids (mmol/l)
LDL-cholesterol3.1 (2.6; 3.6)3.0 (2.6; 3.6)3.0 (2.5; 3.6)2.9 (2.6; 3.5)
HDL-cholesterol1.6 (1.3; 1.9)1.2 (1.0; 1.4)1.3 (1.0; 1.5)1.0 (0.9; 1.2)
Triglycerides0.6 (0.4; 0.9)1.0 (0.7; 1.5)0.9 (0.7; 1.2)1.4 (0.9; 1.9)
Blood glucose (mmol/l)4.6 (4.4; 4.9)5.00 (4.65; 5.45)4.85 (4.60; 5.20)5.20 (4.80; 5.80)
Blood insulin (mIU/l)4.4 (3.2; 6.3)10.6 (7.7; 17.3)9.5 (7.2; 13.4)12.9 (9.1; 21.0)
HOMA-index0.9 (0.7; 1.3)2.1 (1.6; 3.2)2.0 (1.5; 2.8)2.6 (1.8; 4.1)
Blood creatinine (μmol/l)60.0 (54.0; 68.0)61.0 (54.0; 69.0)60.0 (52.5; 67.5)62.0 (56.0; 70.0)
C-reactive protein (mg/)1.45 (1.00; 2.45)8.00 (5.00; 13.00)8.00 (5.00; 13.00)8.00 (4.50; 13.50)
Post-menopausal, n (%)7/57 (12.3)18 (11.9)12 (13.5)6 (9.7)
Platelets (Giga/l)252 (207; 281)277 (239; 309)275 (233; 310)280 (239; 309)
Adiponectin (ng/ml)11,7 (9,5; 13,9) (n = 43)6,8 (5,2; 8,4) (n = 27)7,1 (5,0; 8,3) (n = 14)5,8 (5,6; 8,4) (n = 13)

Comparisons of MPs between obese and non-obese women are shown in Table 2.

Table 2. Comparison of micropartiles (MPs) and leukocyte-platelet aggregates between obese and non-obese women
 Non-obese (n = 60)Obese (n = 151)Unadjusted P valueAdjusted P valuec
  1. Data are presented as median and range between 25th and 75th percentile.
  2. aP value from Mann–Whitney test; NS: non significant.
  3. bP value from Student t test.
  4. Multivariate regression was used to estimate the association between each type of MP or aggregates and obesity.
  5. cAdjustments were made for smoking, blood pressure, MS, hs-CRP concentration, HOMA, adiponectin, age, platelet, neutrophil and monocyte number.
Total MPs (n/μl)668.0 (388.0; 1083.0)1563.0 (875.0; 2865.0)<0.001a0.01
Platelet MPs (n/μl)456.0 (277.0; 759.0)1060.0 (558.0; 2197.0)<0.001a<0.01
Endothelial MPs (n/μl)8.0 (6.5; 9.5)14.5 (10.0; 17.0)<0.001b<0.01
Neutrophil-platelet aggregates (Giga/l)0.430 (0.330; 0.590)0.592 (0.504; 0.787)<0.001a<0.01
Monocyte-platelet aggregates (Giga/l)0.060 (0.040; 0.080)0.077 (0.062; 0.095)0.005aNS

Total and PMPs were markedly and significantly increased in obese as compared to lean women (P < 0.001, univariate analysis). In univariate analysis, significant associations were also seen between total and PMPs and platelet number. Multivariate regression analysis demonstrated a persistent relationship between both types of MPs and obesity (P < 0.01) after adjustment for smoking, blood pressure, MS, hs-CRP concentration, HOMA index, adiponectin, age, platelet, neutrophil, and monocyte count. Similarly, EMPs were significantly increased in obese versus non-obese subjects (Table 2, P < 0.001). This increase was independent of all studied factors. Neutrophil- and monocyte–platelet aggregates were slightly but significantly increased in obese versus non-obese subjects. Concerning monocyte–platelet aggregates, the effect did not persist after adjustment on platelet and monocyte number. By contrast, the level of procoagulant MPs was similar (7 nmol/l) between obese and non-obese women (Table 2). This result was not modified by multivariate analysis.

MPs, LPAs, and metabolic status of obese subjects

Within the obese group, MPs and LPAs were compared according to the metabolic status of the patients (Table 3). In 62 of them (41%), MS was present. There was no statistical difference between the two groups (with and without MS). These results suggest that metabolic factors have no impact on MPs and platelet–leukocyte aggregates.

Table 3. Concentrations of micropartiles (MPs) and leukocyte-platelet aggregates in obese subjects without and with metabolic syndrome (MS)
 Obese without MS n = 89Obese with MS n = 62P value
  1. MPs: microparticles; NPAs: neutrophil–platelet aggregates; MPAs: monocytes–platelet aggregates. Comparisons were performed using ANCOVA.
Total MPs (n/μl)1,634.0 (937.0; 2,953.0)1,515.0 (770.0; 2,699.0)0.29
Platelet MPs (n/μl)1,185.0 (629.0; 2,251.0)1,027.0 (473.0; 2,020.0)0.22
Endothelial MPs (n/μl)15.0 (10.0; 16.0)14.0 (10.0; 18.0)0.65
NPA (Giga/l)0.59 (0.51; 0.72)0.62 (0.50; 0.85)0.53
MPA (Giga/l)0.08 (0.06; 0.09)0.08 (0.05; 0.10)0.88

Effect of weight loss

Among the 84 subjects who had a 1-year follow-up after inclusion or surgery, the weight loss was successful in 32; this means that they lost at least 25% of their excess weight. Among them, 27 had bariatric surgery. As shown in Table 4, the 32 subjects in this group had a mean total body weight loss of 24% (from 117 to 89 kg) and a mean BMI reduction of 19% (from 43 to 35 kg/m2) after 1 year. The successful weight loss was associated with a significant improvement of all studied clinical and metabolic parameters including blood pressure, HDL-cholesterol, insulinemia, HOMA index, and C-reactive protein (all P < 0.001). By contrast, no significant improvement in any of these parameters was observed in subjects without successful weight loss (Table 4).

Table 4. Clinical and metabolic characteristics of patients with weight loss, at baseline and one year later
n = 32Baseline1 year laterPPa
  1. MPs: microparticles; NPAs: neutrophil–platelet aggregates; MPAs: monocytes–platelet aggregates. Comparisons were performed using ANCOVA.
  2. aAdjustments were made for smoking, blood pressure, MS, hs-CRP concentration, HOMA, adiponectin, age, platelet, neutrophil and monocyte number.
Period (days)331.4 ± 88.9  
Weight (kg)117 (104; 129)89 (83; 96)<0.001
BMI (kg/m2)43 (41; 50)35 (32; 37)<0.001
Waist to hip ratio0.87 (0.83; 0.93)0.84 (0.80; 0.88)<0.001
Systolic BP (mmHg)122 (120; 140)120 (115; 130)<0.001
Diastolic BP (mmHg)80 (70; 82)70 (60; 70)<0.001
HDL cholesterol (mmol/l)1.2 (0.9; 1.4)1.4 (1.1; 1.6)<0.001
Triglycerides (mmol/l)1.0 (0.8; 1.5)0.8 (0.6; 1.1)<0.001
Blood insulin (mIU/l)11.2 (9.1; 14.8)6.3 (4.6; 10)<0.001 
HOMA index2.2 (1.8; 3.0)1.3 (1.1; 2.2)<0.001
C-reactive protein (mg/l)8.5 (5.7; 11.5)3.0 (2.0; 6.2)<0.001
Platelet MPs (n/μl)1320 (676; 2396)843 (398; 1628)0.120.37
Endothelial MPs (n/μl)15 (9.7; 16.5)18 (9.5; 22)0.360.28
NPA (Giga/l)0.69 (0.57; 0.77)0.59 (0.39; 0.92)0.150.12
MPA (Giga/l)0.09 (0.07; 0.10)0.07 (0.05; 0.13)0.110.09

The median number of PMPs decreased by 36% in the group of subjects who lost weight successfully and by 33% in the other group. This decrease did not reach statistical significance. EMPs increased slightly in both groups (+20 and +30%, respectively) while both types of LPAs did not vary significantly in any group. In addition, among the 84 subjects who had a 1-year follow-up, no correlation was found between MP or LPA levels and BMI (all Spearman's coefficients < 0.3, data not shown).

We also compared MPs and LPAs between obese subjects who had lost weight and non-obese subjects. Our results show that total MPs, PMPs, EMPs, and both types of LPAs remained significantly higher in patients after weight loss than in control patients (P = 0.01; P = 0.01; P < 0.001; P = 0.04, respectively).

Discussion

Our results show that, in a well-defined population of women with severe obesity (mean BMI: 42 kg/m2), platelet and EMPs and NPA are significantly (P < 0.01) increased as compared to lean women of similar age. This increase is independent of classical cardiovascular risk factors either considered individually or clustered as MS. We found no association with CRP and adiponectin levels nor with the WHR. Moreover, within obese subjects, platelet and EMPs as well as LPA levels did not differ significantly between subjects with and without MS.

Three studies have previously evaluated MPs in obesity [11, 12, 21]. Our results are in accordance with those obtained by Murakami et al. and Esposito et al. as they clearly confirm the significant association between platelet as well as EMP levels, and obesity. By contrast, in our study, and at variance with Goichot et al., procoagulant MPs did not differ significantly between obese and non-obese subjects. The reasons for this discrepancy are unclear as MP counts and assay of their procoagulant activity were done using the same plasma samples, thus excluding an artifact in pre-analytical preparation of plasmas. The method used in our study is that described by Hugel et al. [19], now commercialized by Hyphen BioMed®. The capture of MPs is achieved through interaction of phosphatidylserine with immobilized annexin V. Goichot et al., by contrast, used home-made reagents, which may have induced some differences. Differences in both patient and control populations may also account for this discrepancy. In the present study, total MPs were identified using flow cytometry by their binding of fluorescent annexin-V to phospholipids while PMPs also expressed GPIIb, a constitutive platelet membrane protein. PMPs represented 68% of the total MP amount in controls as well as in obese subjects (data not shown). Enhanced vesiculation of other cell types, notably endothelial cells (as shown in the present study) and leukocytes, has been reported in obese and probably accounts for the increase of non-PMPs [21]. The discrepancy between total MPs that are measured via their phosphatidylserine expression and the prothrombinase activity assay of MPs is surprising; yet, similar data have been obtained previously in type 2 diabetes [22]. The authors speculated that a difference in aminophospholipid content could account for this discrepancy.

Our study also shows, for the first time, the significant increase in obese subjects of NPA. Enhanced platelet activation and formation of platelet–leukocyte aggregates have been previously described in acute coronary syndromes and acute ischemic stroke [23]. Enhanced platelet activation has been described in obese patients; activated platelets degranulate and release MPs that express P-selectin at their surface; P-selectin-positive platelets or MPs adhere to leukocytes via their P-selectin ligand, P-selectin glycoprotein ligand-1 (PSGL-1). Platelet–leukocyte interaction induces leukocyte activation and release of inflammatory cytokines [24]. In obese subjects, persistent circulation of activated leukocytes may cause substantial impairment of the cerebral and cardiovascular microcirculation.

In our study, the association between platelet, EMPs, NPA, and obesity appeared to be independent of all the factors, notably the classical cardiovascular risk factors, tested in the multivariate analysis (blood pressure, smoking, estroprogestative treatment, MS, CRP concentration, age, and platelet count). Using the IDF criteria, these factors were clustered as MS in 41% of the obese subjects (n = 62). Using another definition of MS, such as the NCEP-ATPIII score [25], did not change the number of MS+ patients in our population. When MS+/MS− obese subjects were directly compared, we found similar levels of all types of MPs and NPA in the two groups.

These results are in line with those obtained in several studies with obese subjects. Arteaga et al. [26] compared 33 patients with MS (all of them having a BMI > 29.9 kg/m2) to 25 controls without MS (mean BMI of 27.6). Free PMPs were not higher in the MS group as compared to the non-MS group. Interestingly, in a recent study [27], the adipose tissue from obese individuals without metabolic complications was shown to secrete cytokines altering the endothelium function through activation of the NF-KB pathway. Similarly, in obese children MP release was found to correlate only with BMI [28]. Our results are also in accordance with previous studies as the population of obese subjects studied by Goichot et al. was devoid of any cardiovascular risk factor [12].

By contrast, in non-obese subjects, total or subtypes of MPs have been associated with MS criteria [29], notably diabetes and hypertension [4]. In 467 healthy non-obese Japanese subjects, without a history of cardio- or cerebrovascular disease, platelet-derived MPs identified by CD42b and CD42a (glycoprotein Ib and IX) were positively associated with the MS [29]. In 216 subjects without cardiovascular disease and subnormal BMI (26 ± 4 kg/m2), Chironi et al. reported that the level of leukocyte-MPs but n either EMPs n or PMPs was significantly elevated in those with MS [31]. Moreover, circulating MPs are increased in most cardiovascular diseases and preliminary data indicated that plasma levels of MPs could be of prognostic value for the occurrence of cardiovascular diseases, notably the occurrence of secondary myocardial infarction or death in patients with acute coronary syndromes [32]. However, while elevated numbers of MPs are likely to reflect the cellular stress of endothelial cells and platelets and thus contribute to the progression of endothelial damage, recent studies suggested that MPs may also have a protective effect [7, 33].

In our study, the observed effect of successful weight loss on MP levels was unexpected as, in spite of a significant beneficial effect on all metabolic abnormalities, EMPs, PMPs, and LPAs did not vary significantly. This reinforces the lack of correlation between metabolic abnormalities and these variables. For the purposes of our study, we defined successful weight loss as that representing a loss of at least 25% of excess body weight at one year. This constitutes one limitation of our study; this choice takes into account the heterogeneity of the patients, 34 of whom underwent bariatric surgery and 50 of whom only had medical follow-up. The median BMI, after a successful weight loss defined as more than 25% of the excess weight, was still 35 kg/m2. As surgical patients were still losing weight at one year, these results suggest that a longer follow-up would be of importance to show whether a larger weight loss is required to significantly impact MP and LPA levels, or if these markers would remain high despite being lower in patients who are not obese. Incidentally, in the study by Murakami et al. [11], a mean weight loss of 8 kg in moderately obese subjects (mean BMI ± SEM = 27.4 ± 0.3 kg/m2), was associated with the decrease of circulating PMP. As in our study, in a recent work by Morel et al. [34], a short-term very low caloric diet had no effect on PMPs nor EMPs determined using flow cytometry. The method used to quantify MPs is of importance as in both studies PMPs did not vary when assayed using flow cytometry, while procoagulant-platelet derived MP specifically decreased in the study by Morel et al. This was accompanied by a reduction in leptin levels. The lack of correlation between leptin and platelet-MPs did not substantiate the view that leptin directly modulate platelet activation [34].

Moreover, gastric bypass is known to produce significant improvement in glucose homeostasis via incretin variations [35], and glucose homeostasis can influence MP levels. In our successful weight loss group, 27 patients underwent bariatric surgery (9 patients had gastric banding, 18 had gastric bypass) and 5 patients had medical follow-up. Comparing the MP differences (before versus after one-year follow-up) between these subgroups, we found no statistically significant difference (data not shown); this result does not suggest a specific role of the incretin variations induced by gastric bypass in MP levels. MPs have pro-thrombotic and pro-inflammatory effects, processes which are involved in the physiopathology of cardiovascular diseases. The MHO individuals are insulin sensitive and have normal lipid profiles, despite having excessive fatness. Although a matter of controversy, the recent report of the American Heart Association [3] clearly indicates that obesity is associated with an increased risk of cardiovascular mortality and morbidity and that this risk is independent of classical cardiovascular risk factors. Our results further underline that metabolic abnormalities are not involved in MP or LPA increase in obese.

Recently, several markers including inflammatory cytokines (TNF, IL-6), metalloproteinases (MMP-2 and MMP-9), adipokines (leptin, adiponectin), lipoprotein-phospholipase A2, and soluble CD40 ligand [36] have been proposed to better evaluate the cardiovascular risk in obese subjects. The effect of weight loss on these traditional and non-traditional cardiovascular risk factors has been evaluated in several studies with controversial results as concerns notably the inflammatory profile and MMP-9 levels [36, 37]. By contrast, the recovery from “adiposopathy,” characterized by improved adiponectin and decreased leptin concentration occurs constantly after bariatric surgery [38]. Adiponectin is an adipose-tissue-derived peptide that is believed to have significant anti-atherogenic and anti-thrombotic properties. Interestingly, in a recent study, Wolk et al. [10] found that patients with acute coronary syndrome (ACS) had significantly lower adiponectin levels than those without ACS, independent of a variety of cardiovascular risk factors. Adiponectin was therefore proposed as an obesity biomarker to define an “obesity phenotype” relevant to cardiovascular disease. In the present study, no association was found between MPs and adiponectin concentration. However, the dosage was done in a limited number of subjects. Hyperleptinemia activates platelets and affects endothelial nitric oxide production [39]. However, our results suggest that a parameter not responding to weight loss is involved in the activation of platelets, endothelium, and leukocytes, but this is at the one-year time point. Leptin, which decreases markedly with weight loss, is not a good candidate as previously suggested by Morel et al. Lipoprotein-phospholipase A2 and/or soluble CD40 ligand might be good candidates which need to be evaluated.

Whether increased MPs specifically reflect an increased cardiovascular risk in those obese subjects who are free of metabolic abnormalities is an interesting question which should be evaluated in a future study. Numerous studies also associated MPs with a prothrombotic context, emphasizing their significance as a relevant parameter in the identification of patients at higher risk of venous thrombosis.

Alternatively, it is now known that MPs, including EMPs, may also have beneficial effects protecting cells against the consequences of external stimuli or stress [7, 40]. Because of their pleiotropic effect, it will be of utmost interest to evaluate the prognosis significance of platelet and EMPs and of LPA in obese patients in the long term. Prospective studies are further needed to clarify this point.

Acknowledgments

Authors sincerely thank Dr. Annick Dupré and Dr. France Teillet (hôpital Louis Mourier) who enrolled the control subjects, Dr. Catherine Bogard (hôpital Louis Mourier) who supervised the biochemistry assays, Dr. Anne Marie Vissac and Jean Amiral (HYPHEN BioMed) for giving them the Zymuphen MP-Activity kits, the nurses Nadine Martinez and Denise André for their dedication to this project, Eugène Fammègne and Maryse Lamotte for taking care of the samples of cohort Colombes subjects, Brigitte Martel and Anne Marie Duprat who performed the MP assays. They also greatly acknowledge the patients and controls of cohort COLOMBES who agreed to participate to the study, and the Centre de Ressources Biologiques—Colombes of Louis Mourier Hospital for the conservation of plasma samples.

Ancillary