Plasma Phospholipid Transfer Protein Activity Is Decreased in Type 2 Diabetes During Treatment With Atorvastatin

RESEARCH DESIGN AND METHODS

This study comprised 215 patients enrolled in the DALI study. DALI was a double-blind randomized placebo-controlled multicenter study evaluating the effect of 10 vs. 80 mg atorvastatin daily on lipid metabolism, endothelial function, coagulation, and inflammatory factors in unrelated men and women with type 2 diabetes. The protocol and eligibility criteria have been described in detail elsewhere (35). Briefly, men and women aged 45–75 years with duration of diabetes for at least 1 year and an HbA_1c (A1C) ≤10% were eligible. The diagnosis of type 2 diabetes was defined according to the American Diabetes Association classification (36). Lipid inclusion criteria were total cholesterol between 4.0 and 8.0 mmol/l and fasting triglycerides between 1.5 and 6.0 mmol/l. Patients were recruited in Leiden, Rotterdam, and Utrecht, the Netherlands. The ethical committees of the participating centers approved the study protocol, and written informed consent was obtained from all subjects.

Laboratory measurements.

After an overnight fast for a minimum of 12 h, blood was drawn into EDTA-containing tubes for analysis of lipid profiles at baseline and after treatment for 30 weeks. Plasma was prepared by immediate centrifugation (3,000g, 15 min, 4°C), and samples were stored at −80°C for further analyses. Cholesterol and triglycerides were determined by enzymatic colorimetric methods on a Hitachi 911 automatic analyzer (Boehringer Mannheim, Mannheim, Germany). ApoB and apoAI were assayed, using automated immunoturbidimetric assays (Tina-Quant; Roche Diagnostics, Mannheim, Germany). Plasma apoE was analyzed automatically with a nephelometric assay (Wako, Osaka, Japan). Plasma HDL cholesterol was measured by a direct enzymatic method based on polyethylene glycol-modified enzymes, using a Hitachi 911 autoanalyzer (Roche Diagnostics). LDL cholesterol was estimated with the Friedewald formula (37). Fasting plasma glucose was determined on a Hitachi 917 analyzer using an UV-hexokinase method (catalog no. 18766899; Boehringer Mannheim). A1C was determined by high-pressure liquid chromatography, using the Variant method (catalog no. 270-0003; BioRad).

PLTP mass and activity.

Plasma PLTP mass was analyzed as previously described (29). The between-assay coefficient of variation (CV) and within-assay CV were <3.5 and <2.8%, respectively. Plasma PLTP activity was measured in a liposome vesicles-HDL system as described previously (38,39). In short, plasma samples were incubated for 45 min, 37°C, with [¹⁴C]dipalmitoyl-phosphatidylcholine-labeled liposomes and an excess of HDL, followed by precipitation of the liposomes with a mixture of NaCl, MgCl₂, and heparin (final concentrations 230 mmol/l, 92 mmol/l, and 200 units/ml, respectively). The measured PLTP activities are linearly related to the amount of plasma used in the incubations (0.5–1.0 μl). The method is not influenced by the phospholipid transfer-promoting capacities of cholesterylester transfer protein (38). Plasma PLTP activity was related to the activity measured in a reference plasma pool, based on plasma collected from 200 randomly selected healthy individuals, and is expressed as the percentage of the activity measured in this reference plasma, which was included in every assay. The PLTP activity in this reference plasma was 21.2 μmol · ml⁻¹ · h⁻¹. The between- and within-assay CVs were 4.8 and 3.5%, respectively.

Statistical analysis.

All data are expressed as the means ± SD. All statistical analyses were carried out with SPSS software (version 12.0). Mean differences between the groups were analyzed, using ANOVA. To analyze determinants of PLTP activity and PLTP mass, we used the following strategy. First, univariate linear regression analyses was performed in the total cohort with PLTP activity and PLTP mass as dependent variables and the determinant of interest as an independent variable. Second, multivariate regression analysis was performed with PLTP activity and PLTP mass as dependent outcome variables, sex and waist as confounders, and statistically significant variables as independent variables. Third, the effect of treatment on PLTP activity, PLTP mass, and apoE was analyzed with two-way ANCOVA, using a placebo-controlled setting. Throughout, a two-tailed P value of <0.05 was interpreted as indicating a statistically significant difference.

RESULTS

The characteristics of the 215 patients have been extensively described (35) and are summarized in Table 1. The atherogenic lipoprotein profile of the patients was characterized by elevated plasma triglycerides (2.81 ± 1.01 mmol/l) and decreased plasma HDL cholesterol (1.04 ± 0.24 mmol/l) levels. The mean plasma cholesterol levels were 6.0 ± 0.9 mmol/l. Average A1C was 8.3 ± 1.1%, and plasma glucose was 10.6 ± 3.2 mmol/l. The average duration of diabetes was 10.5 years.

Baseline findings.

Baseline plasma PLTP mass concentrations varied from 2.50 to 24.50 mg/l, with an average of 8.84 ± 3.01 mg/l, and were similar in the three treatment groups (Table 2). Plasma PLTP activity at baseline varied from 64.4 to 187.1%, with an average of 107.6 ± 19.4%, and was identical in the three treatment groups. The correlation between baseline plasma PLTP mass and activity was low (r = 0.120, NS). Baseline plasma apoE concentrations were 4.49 ± 1.02 mg per 100 ml. One important determinant for plasma apoE levels is the apoE genotype. In the present study cohort, the apoE genotypes were in Hardy-Weinberg equilibrium (χ² = 2.916, degrees of freedom = 3, P = 0.4): 62% of the patients had the apoE3/E3 genotype, 17% apoE2/E3, 3% apoE2/E4, 17% apoE3/apoE4, and 1% apoE4/E4. The average apoE and PLTP activity levels per apoE genotype group are shown in Table 3. No significant effect of apoE genotype on PLTP activity was found.

Waist circumference, plasma glucose, A1C, and plasma apoE were associated with baseline PLTP activity (Table 4, Fig. 1). In multivariate analyses, adjusted for age and sex, only waist circumference and plasma apoE remained major determinants of plasma PLTP activity. None of the other lipid (including cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides) and apo (apoAI and apoB) parameters showed a significant correlation. Plasma PLTP mass is known to be predominantly determined by the levels of plasma HDL cholesterol and apoAI (21). In our analyses both HDL cholesterol and apoAI contributed to the variation of plasma PLTP mass, independent of age, sex, and waist circumference. Plasma apoE and triglyceride concentrations were less important determinants (Table 5).

Treatment with atorvastatin.

As expected, adjusted for the placebo effect, treatment with atorvastatin resulted in a dose-dependent decrease in plasma total cholesterol, LDL cholesterol, and apoB (all P < 0.001) (35). Plasma HDL cholesterol increased by 6.0 and 5.2% with 10 and 80 mg atorvastatin, respectively (P < 0.05), whereas plasma apoAI levels remained unchanged (Table 2). Plasma PLTP activity decreased by 8.3% (P < 0.01) and 12.1% (P < 0.001), respectively. The difference between the two groups did not reach statistical significance, but the trend was significant (P < 0.01). In contrast, plasma PLTP mass increased by 2.8% (NS) and 11.6% (P < 0.001), respectively. The difference between 10 and 80 mg atorvastatin was significant (P < 0.005). Plasma apoE levels were strongly decreased after atorvastatin treatment by 28 and 36%, respectively (both P < 0.001 vs. control). In a multivariate placebo-controlled analyses, including all lipid and apo parameters, the decrease in PLTP activity was only predicted by the decrease in apoE (standardized coefficient β = 0.529, P < 0.0001) (Fig. 2). Similar results were obtained when only individuals with the apoE3/E3 genotype were analyzed (n = 133, r = 0.513, P < 0.001), illustrating that the effect is independent of the apoE genotype.

DISCUSSION

In patients with type 2 diabetes, we found that atorvastatin treatment resulted in a decrease in PLTP activity and an increase in PLTP mass, leading to a substantial change in mass-adjusted activity. Statistical analysis revealed that the change in apoE and the change in PLTP activity are associated, thereby supporting the concept that apoE may be essential for the stability of the PLTP protein and its activity.

It has been shown that patients with type 2 diabetes have increased levels of plasma PLTP activity (19,40). Positive associations have been reported for PLTP activity and parameters of body composition, such as body fat, BMI, and waist circumference (22,23,41). Also in the present study, both waist circumference and BMI were positively correlated with PLTP activity. In multivariate analysis, however, waist circumference was the better predictor. Increase in abdominal fat accumulation, as represented by increase in waist circumference, strongly relates to insulin resistance (42). Therefore, it is not surprising that patients with hypertriglyceridemia without diabetes have increased levels of PLTP activity as well (21). However, neither plasma glucose levels nor A1C were strong predictors for PLTP activity in the present study, even though glycemic control was suboptimal. Also, no sex-specific difference in PLTP activity or mass was found.

Mice with a genetic deficiency of PLTP have impaired hepatic secretion of apoB-containing particles, which is in favor of a role for PLTP in the biosynthetic pathway of VLDL by enabling donation of lipids to the nascent particles (14). In accordance with this mechanism, overexpression of human PLTP in transgenic mice results in elevated rates of hepatic VLDL secretion (43). Indeed, a number of clinical studies have reported the existence of a link between plasma apoB or LDL cholesterol and PLTP activity (20,41), but in the present study, we were unable to find evidence for this association. It has to be noted that we selected our patients on the presence of raised plasma triglycerides. As a result, we observed relatively small ranges in plasma apoB and cholesterol levels (Table 1), which could have obscured a relationship between PLTP activity and LDL cholesterol levels. Positive correlations with plasma triglycerides and cholesterol have been reported in a Finnish population study (44), but this could not be replicated in the present study, where correlations with triglycerides (r = 0.10, NS) and cholesterol (r = 0.00, NS) were only weak or absent.

Plasma apoE levels at baseline were associated with PLTP activity in both univariate and multivariate analyses. None of the other lipid variables or determinants of glucose homeostasis, like glucose levels and A1C, showed any significant contribution. PLTP mass, on the other hand, strongly positively correlated with plasma HDL cholesterol and apoAI and negatively with plasma triglycerides in univariate analysis. Multivariate analysis revealed that PLTP mass was predominantly explained by HDL cholesterol levels. Plasma PLTP is present in two forms: one form with little, if any, phospholipid transfer activity and another form with high activity. Both are associated with macromolecular complexes of different size (30,32). The inactive form is present on larger HDL particles, compared with the active form. The currently used PLTP mass assay has been extensively validated and is capable of analyzing both the catalytically active and inactive forms (45). Because the greater part of plasma PLTP is inactive and because inactive PLTP copurifies with apoAI (32), it can be appreciated that PLTP mass and apoAI, as well as HDL cholesterol, are related (21).

Atorvastatin treatment significantly reduced plasma apoE levels and plasma PLTP activity, whereas total PLTP mass was increased. Evidence for a possible role for apoE in PLTP homeostasis has been obtained in a number of different studies. The active form of PLTP copurifies with apoE, whereas the inactive form is associated with apoAI (32). In addition, active PLTP secreted from HepG2 cells coelutes with apoE but not with apoAI (46). Furthermore, apoE but not apoAI is able to convert inactive PLTP into the active form (34). In light of these and our present observations, we hypothesize that atorvastatin, via its lowering effect on apoE concentration mediated through elevation of hepatic LDL receptors, destabilizes the active form of PLTP. As a consequence, PLTP could be partly inactivated, resulting in lowering of PLTP activity. Finally, more inactive PLTP may have contributed to the observed elevation of plasma PLTP mass. In conclusion, the present study shows, for the first time, that a statin is able to decrease PLTP activity and supports our hypothesis that apoE may stabilize PLTP in vivo.

APPENDIX

The DALI Study Group.

In alphabetical order: Erasmus Medical Center Rotterdam, Department of Internal Medicine: I. Berk-Planken, N. Hoogerbrugge, and H. Jansen; Erasmus University Rotterdam, Departments of Biochemistry and Clinical Chemistry: H. Jansen; Gaubius Laboratory TNO-KvL, Leiden: H.M.G. Princen; Leiden University Medical Center: M.V. Huisman and M.A. van de Ree; University Medical Center Utrecht, Julius Center for General Practice and Patient Oriented Research: R.P. Stolk and F.V. van Venrooij; University Medical Center Utrecht, Division of Internal Medicine: J.D. Banga, G.M. Dallinga-Thie, and F.V. van Venrooij.