Pravastatin reduces fibrinogen receptor gpIIIa on platelet-derived microparticles in patients with type 2 diabetes


Dirkje W. Sommeijer, Laboratory for Experimental Internal Medicine, G2-108, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands.
Tel.: +31-20-5665985; fax: +31-20-6977192; e-mail:


Diabetes mellitus type 2 is characterized by multiple cardiovascular risk factors including hypercholesterolemia [1], inflammation [2] and activated coagulation [3]. An increased level of circulating microparticles (MP) was suggested to be one of the procoagulant determinants in patients with type 2 diabetes that may contribute to the high risk of atherothrombotic events [4]. Treatment with hydroxymethyl-glutaryl-CoA reductase inhibitor (HMG-CoA) reductase inhibitors (known as statins) reduces the risk of cardiovascular complications by 25% in patients with or without type 2 diabetes [5,6]. There is emerging evidence that the beneficial effects of statins also involve effects on platelet activation and aggregation [7]. It is not known whether statins influence the number and properties of MP, and in particular of platelet-derived MP, in patients with type 2 diabetes. The two objectives of the present study were: (1) to assess the effect of pravastatin on the level and cellular origin of MP and (2) to evaluate its effect on the composition of MP.

A group of 50 patients with type 2 diabetes was studied in a cross-over trial, described in detail elsewhere [8]. Plasma samples for MP evaluation were available from 48 patients. Men and women between 18 and 80 years of age with type 2 diabetes for at least 1 year and serum cholesterol levels of 5.0–10.0 mmol L−1 were eligible for the study. An open-label randomized cross-over design was used. One half of the subjects started out with pravastatin (Selektine; Bristol Myers Squibb, The Netherlands) 40 mg day−1 for 8 weeks, while the other half received no treatment. Laboratory outcomes at 8 and 16 weeks were compared, each patient being his/her own control. Ten subjects without type 2 diabetes were enrolled from the outpatient eye clinic and from the personnel of the Slotervaart Hospital and served as a healthy control group. All patients gave informed consent and the study was approved by the institutional Ethical Review Board of the Slotervaart Hospital, Amsterdam.

The antibodies used have been described elsewhere [9], with the following exceptions: CD62e-PE (clone HAE-1f, IgG1) (Ancell; Bayport, MN, USA), CD54-PE (clone 84H10, IgG1), CD62p-PE (clone CLB-Thromb/6, IgG1) and CD144-FITC (rabbit polyclonal) from MedSystems (Vienna, Austria). Annexin V-APC was from Caltag Laboratories (Burlingame, CA, USA), CD61-PE (clone VI-PL2, IgG1,κ) and CD15-PE (clone HI98, IgMk) from Pharmingen (San Jose, CA, USA). CD4-PE (clone CLB-T4/2, 6D10, IgG1) and CD66e-PE (clone CLB-gran/10, IH4Fc, IgG1) were obtained from CLB (Amsterdam, The Netherlands). CD106-FITC (clone 1.G11B1, IgG1) was from Calbiochem (Darmstadt, Germany). Blood samples were obtained by standard venepuncture between 9 and 11 am, after 12 h of fasting. MP isolation was performed as described [9]. For flow cytometry, 25 μL MP suspension was diluted with 85 μL phoaphate-buffered saline/citrate buffer, of which 5 μL was used per incubation with MoAb and annexin V. The samples were analysed in a FACSCalibur flow cytometer with CellQuest-PRO software (Becton Dickinson, San Jose, CA, USA) as described [9–11]. To estimate the total number of MP (×106 L−1), the number of MP (N) identified by forward scatter (FSC), sideward scatter (SSC) and binding to annexin V was used in the formula: Number (×106 L−1) = N × [100/5] × [355/70] × [250] in which 5 (μL) is the volume of MP suspension, 100 is the total volume of washed MP suspension, 355 is the total volume in the tube before analysis, 70 is the sample volume analysed and 250 is the original volume of plasma. To estimate the marker positive number of MP (×106 L−1), the number of MP (N) found in the upper right (marker positive and tissue factor (TF) positive) and lower right (marker-positive and TF-negative) quadrants of the flow cytometry analysis (FL1 vs. FL2, corrected for isotype control and autofluorescence) was used in the same formula. Continuous variables are represented as median and 25–75% quartiles. Statistical differences between before and after pravastatin treatment were tested by Wilcoxon rank test. Statistical differences between type 2 diabetes patients and controls were tested by Mann–Whitney U-test. A P-value of <0.05 indicated a significant difference. All computations were performed using SPSS 11.0.

The median age of the 48 patients was 59 years. Patients were overweight [median body mass index (BMI) of 29 kg m−2], but well controlled for their diabetes (median glycosylated haemoglobin (HbA1c) 6.9%). For the entire group, lipid profiles and medication were previously reported [8]. Thirteen patients used acetylsalicylic acid in a dosage of 38 or 100 mg day−1 which they kept using during the whole study period. The median age of the 10 controls was similar to that of the patients with type 2 diabetes, 60 years, while BMI (median 26.3 kg m−2), total and low-density lipoprotein (LDL) cholesterol were all significantly lower in the controls compared with the patients with type 2 diabetes.

After treatment with pravastatin, a statistically significant reduction of 22% of total cholesterol [−1.4 mmol l−1 (−1.9, −1.0)], a 32% reduction of LDL cholesterol [−1.3 mmol l−1 (−1.74, −0.95) (median reduction and 27–75% percentiles)] and a 10% reduction of triglycerides [−0.19 mmol l−1 (−0.55, 0.08)] was observed in the patients. High-density lipoprotein (HDL) cholesterol levels did not change during treatment. In patients with type 2 diabetes (n = 48) and controls (n = 10), the total numbers of annexin-positive MP and annexin-TF-positive MP were not different [between controls and patients: 523 (377–614) × 106 L−1 vs. 434 (327–591) × 106 L−1; P = 0.3] (Fig. 1 upper panel A and B). However, the number of TF-positive MP was significantly lower in controls compared with patients [6 (5–16) × 106 L−1 vs. 17 (9–24) × 106 L−1; P = 0.03]. MP number before and after pravastatin treatment remained unchanged [434 (327–591) × 106 L−1 vs. 446 (315–595) × 106 L−1; P = 0.9]. Moreover, the number of TF-positive MP did not change after pravastatin treatment [17 (8–24) ×106 L−1 vs. 16 (10–28) × 106 L−1; P = 0.6].

Figure 1.

Upper panel: (A) Total mean number of annexin-positive MP was similar before and after pravastatin treatment (n = 48) and in controls (n = 10). (B) Total mean number of TF-positive (and annexin-positive) MP did not change after pravastatin treatment. The mean number of TF-positive MP was significantly lower in controls compared to patients with type 2 diabetes (*P = 0.03). Middle panel: (A) The mean number of platelet-derived MP remained unchanged during pravastatin treatment (P = 1.0) (n = 48). (B) Pravastatin treatment had neither a significant effect on the mean number of MP-derived from T-helper cells (CD4), T-suppressor cells (CD8), monocytes (CD14), B cells (CD20), endothelial cells (CD62e), erythrocytes (glyco-A) and granulocytes (CD66b) (P > 0.05) (n = 20). Lower panel: Antigen level on platelet-derived MP was measured in patients with type 2 diabetes (n = 48) before (control) and after (treatment) pravastatin treatment. Mean fluorescence of CD61 antigen (IIIa) reduced significantly after treatment (*P = 0.004). Annexin V and TF exposure did not differ between before and after treatment (P = 0.7 and P = 0.4, respectively).

The effect of pravastatin on the number of MP derived from different cell types was assessed in a subgroup of 20 patients (Fig. 1, middle panel). MP from platelets (CD61) constituted the largest proportion of total MP, 68 (60–77) percent. The number of MP derived from platelets, T-helper (CD4) and T-suppressor (CD8) lymphocytes, monocytes (CD14), B lymphocytes (CD20), granulocytes (CD66b), erythrocytes (glyco-A) and endothelial cells (CD62e) remained unchanged after pravastatin treatment in accordance with total MP number. Moreover, the number of CD54-, CD62p-, CD106-, CD144- (endothelial cells) and CD66e- (epithelial cells and granulocytes) positive MP was similar before and after pravastatin treatment (data not shown). Correlation studies were done between number of total as well as cell-specific MP and cholesterol levels. The total number of annexin V-positive MP did not correlate with baseline levels of total cholesterol (P = 0.2), LDL-cholesterol (P = 0.3), HDL-cholesterol (P = 0.5) or triglycerides (P = 0.4). From the subgroups of cell-specific MP, only the number of P-selectin-positive MP (CD62p) correlated with total cholesterol levels (r = 0.57, P = 0.009). The number of MP from B lymphocytes (CD20) and endothelial cells (CD66p) correlated with LDL-cholesterol levels (r = 0.45; P = 0.05 and r = 0.70; P = 0.001, respectively). MP derived from T-suppressor cells (CD8) correlated with triglyceride levels (r = 0.50; P = 0.03).

To assess whether pravastatin has an effect on the membrane composition of platelet-derived MP, the intensity of annexin V, TF and GPIIIa (determined by exposure of CD61 antigen) staining per MP was measured (Fig. 1, lower panel). GPIIIa expression significantly decreased after pravastatin ]113.2 (98.4–133.5) vs. 106.0 (87.3–126.0) mean fluorescence; P = 0.004]. Annexin V and TF exposure per platelet-derived MP remained unchanged during treatment with pravastatin (P = 0.7 and P = 0.4, respectively).

To identify possible relationships between exposure of TF and GPIIIa on MP and lipid levels correlation coefficients were calculated. From the TF positive MP only the number of MP from granulocyte origin (CD66b) correlated with LDL-cholesterol (r = 0.47, P = 0.04). No correlation was observed between the other subtypes or total number of TF-positive MP with total cholesterol, LDL-cholesterol, HDL-cholesterol and triglycerides levels. Exposure of GPIIIa and TF on platelet-derived MP did not correlate with baseline cholesterol levels either (P = 0.8 and P = 0.4, respectively). As both GPIIIa exposure and lipid levels decreased after pravastatin treatment, we also calculated the correlation between the changes (level after treatment minus level before treatment) of these parameters. Again no correlation between delta GPIIIa exposure and delta total cholesterol, LDL-cholesterol, HDL-cholesterol and triglycerides could be observed.

The collected data indicate that pravastatin clearly reduced cholesterol levels in the diabetic patients, but did not alter the total number of MP after 8 weeks of treatment. In addition, the number of TF-positive MP and the level of TF antigen per platelet-derived MP was not influenced by drug treatment. Pravastatin treatment, however, significantly reduced GPIIIa antigen on MP from platelet origin. As GPIIIa associates with GPIIb to form the platelet fibrinogen receptor GPIIb/IIIa, reduction of GPIIIa exposure on MP may be a new aspect of the non-cholesterol-lowering effects of pravastatin.

Microparticle composition depends on the cellular origin and the cellular processes triggering their formation [12]. The observed reduction of GPIIIa exposure on platelet-derived MP is therefore indicative of the activation state of platelets. The GPIIb/IIIa receptor is the main platelet receptor for fibrinogen and crucial for thrombus formation. GPIIb/IIIa is released from a storage pool and transported to the cell membrane in activated platelets [13]. The density of this receptor on the MP membrane is influenced by the stimulus that has induced the formation of MP, which has an effect on fibrinogen binding [14]. We speculate that pravastatin treatment inhibited platelet activation and thereby reduced the exposure of GPIIb/IIIA on platelet-derived MP. Indeed, there are various studies showing that statin treatment has an effect on platelet activation [15,16] and platelet-membrane composition [17,18]. The exact mechanism is not known but probably involves a change in cholesterol content of platelet intracellular and extracellular membranes, which alters membrane traffic and fluidity, or a reduction in cytosolic calcium [17,18]. In particular, pravastatin was found to reduce the expression of P-selectin [19,20]. As the reduction of GPIIb/IIIa on MP is not correlated with reduction of lipid levels in the present study, the effect seems to be a non-cholesterol-lowering effect of pravastatin. To further explore this new aspect of pravastatin treatment, studies are needed that include measurements of other platelet-activation-dependent molecule levels, such as P-selectin and measurements reflecting the procoagulant or pro-inflammatory activity of MP.


We thank M de Rijk and OA Ternede, research nurses from the Slotervaart hospital for their great help in this investigation. H ten Cate is a Clinical Established Investigator of the Netherlands Heart Foundation. The study costs were defrayed from the material budget provided by the Netherlands Heart Foundation to DW Sommeijer and H ten Cate.