Patients with type 1 diabetes have greatly increased phospholipid transfer protein (PLTP) activity and have an altered HDL subclass distribution. In 195 patients with type 1 diabetes and in 194 men and women aged 30–55 years, we examined the relationship of PLTP activity to HDL and examined whether PLTP activity contributes to differences in HDL found in type 1 diabetes. PLTP activity was measured using an exogenous substrate assay. Average HDL particle size and HDL subclasses were measured using nuclear magnetic resonance spectroscopy. Apolipoprotein AI (apoAI) and apoAII were measured by immunoturbidimetry. The amount of apoAI present in LpAI was measured using a differential electroimmunoassay, and the amount of apoAI in LpAIAII was inferred from the apoAI and LpAI data. Higher PLTP activity was associated with more large HDL (P < 0.001) and less small HDL (P < 0.01), more apoAI and apoAII (both at P < 0.001), and more apoAI in both LpAI and LpAIAII (P = 0.02 and P < 0.001, respectively). These associations were independent of other lipids and enzyme activities. Adjusting for PLTP activity halved the difference between subjects with and without diabetes in apoA1 (from 10.1 mg/dl higher in subjects with diabetes to 4.6 mg/dl higher) and large HDL (2.4 μmol/l higher to 1.2 μmol/l higher) and reduced the difference in HDL size (from 0.31 nm higher to 0.26 nm higher). PLTP activity was also positively associated with apoB, total VLDL and LDL particle number, and IDL level in subjects with diabetes. These data support the idea that PLTP is a major factor in HDL conversion and remodeling in humans and that higher PLTP activity makes an important contribution to the higher apoAI levels and altered HDL subclass distribution in type 1 diabetes. They also support a role for PLTP in the metabolism of apoB-containing lipoproteins.

HDL exists as several subpopulations that differ in size, density, and apolipoprotein (apo) content as well as capacity to accept cellular cholesterol. These subpopulations undergo remodeling and interconversion by the addition and removal of their neutral lipids, phospholipids, and apo components (1). Proteins known to be involved in HDL metabolism include lecithin:cholesterol acyltransferase (LCAT), cholesterol ester transfer protein (CETP), and the endothelial- bound enzymes hepatic and lipoprotein lipase. Phospholipid transfer protein (PLTP) enhances the transfer of phospholipids from VLDL to HDL particles. In vitro and animal studies also demonstrate a role for PLTP in the remodeling of HDL, including promotion of the generation of pre-β HDL particles and large HDL particles from HDL3 (2–4). However, in vivo data on the relationship of PLTP to HDL metabolism in humans are sparse and in parts conflicting (5–7).

We previously demonstrated that patients with type 1 diabetes have a significantly elevated PLTP activity (∼1 SD higher than in the general population) and that PLTP activity correlates with HDL cholesterol (8). Many studies have demonstrated that, in the absence of renal failure, HDL cholesterol and apoAI are often higher in subjects with type 1 diabetes than in the background population (9). We recently demonstrated that patients with type 1 diabetes have more large and less small HDL with a consequently increased average HDL particle size (10). The role of PLTP activity in these diabetic-nondiabetic differences in HDL has not been examined.

To provide more detailed data on the relationship of PLTP and HDL metabolism in the general population and in patients with type 1 diabetes, we examined the association of PLTP activity with several aspects of HDL, namely, nuclear magnetic resonance (NMR) spectroscopy-defined HDL subclass levels and average particle size, the levels of apoAI and apoAII, and the amount of apoAI that resides in LpAI (HDL particles that contain only apoAI) and in LpAIAII (particles that contain apoAI and apoAII). We examined to what extent the elevated PLTP activity in type 1 diabetes we have recently reported (8) might account for differences in various aspects of HDL between subjects with and without diabetes. As recent studies have suggested a role for PLTP in apoB secretion (11), we also report on the association of PLTP activity with apoB and apoB-containing lipoproteins.

Subjects.

The methods have been described in detail previously (10). A random sample of men and women with type 1 diabetes aged 30–55 years was taken from the diabetes registers of five London Hospitals. Type 1 diabetes was defined by age of onset ≤25 years and onset of permanent insulin therapy within 1 year of diagnosis. A random sample of the general population, stratified to have a similar age and sex distribution to the patients, was drawn from the lists of two London general practices. Patients on renal replacement therapy were excluded. Ethics Committee approval was obtained from our Institution and the Local Research Ethics Committees covering the practices involved. Written informed consent was obtained from all subjects. In all, 198 subjects with and 198 subjects without diabetes were recruited and had PLTP activity measured. Three subjects who had diabetes and were on lipid-lowering therapy and four subjects who did not have diabetes but had hypertriglyceridemia (>6 mmol/l) were excluded from these analyses.

Laboratory methods.

After an overnight fast, blood samples were taken and total cholesterol, HDL cholesterol, and triglyceride levels were measured using enzymatic colorimetric methods (12,13). HDL cholesterol was measured directly after stabilization of other lipoproteins. (14) LDL cholesterol was calculated as described by Friedewald et al. (Friedewald WT, Levy RI, Frederickson DS: Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 18:499–502, 1972). HbA1c was measured using a latex enhanced immunoassay (between-assay coefficient of variation [CV] 1.7%). ApoB was measured by using an immunochemical assay (Orion Diagnostica).

Fasting citrated samples were centrifuged at 4°C within 1 h of sampling and then stored at −70°C. On average, samples had been stored for 8 months before analysis. Plasma PLTP activity was measured with labeled liposome vesicles as phospholipid donor and excess pooled human HDL as phospholipid acceptors, as described. (15). The between-assay CV (i.e., running a reference sample on different days) was 4.8%, and the within-assay CV was 3.0% (i.e., running a reference assay several times within the assay run). Plasma PLTP activity levels were related to the activity in a human reference plasma pool analyzed in each run and are expressed in arbitrary units (AU), corresponding to the percentage of the activity in the reference plasma. CETP activity and the actual mass of cholesterol esters transferred (plasma CET) were measured as previously described (8) (between-assay CV 4.5% and 7.2%, respectively) and modified as described (16). The rate of CET is expressed in absolute units of nanomoles of cholesteryl ester transferred per milliliter of plasma per hour and is constant during 3 h of incubation.

HDL size and subclass concentration were measured using a 400-MHz proton NMR analyzer at Liposcience, Inc., as described (8,17). Subclass particle levels in units of micromoles per liter were measured for five HDL subclasses: H5 (11.5 ± 1.5 nm), H4 (9.4 ± 0.6 nm), H3 (8.5 ± 0.3 nm), H2 (8.0 ± 0.2 nm), and H1 (7.5 ± 0.2 nm). Average HDL particle sizes (nm diameter) were determined by weighting the relative mass percentage of each subclass by its diameter. For simplifying analyses, the HDL subclasses were grouped into three categories: large HDL (H4+H5), intermediate HDL (H3), and small HDL (H1+H2). In-house experiments have shown that the measures are robust to samples that have been frozen. NMR spectroscopy was also used to quantify large VLDL (60–200 nm + chylomicrons), intermediate VLDL (35–60 nm), small VLDL (27–35 nm), intermediate-density lipoprotein (IDL; 23–27 nm), large LDL (19.8–23 nm), and small LDL (18.3–19.7 nm) as well as average VLDL and LDL size as previously described (10).

Serum apoAI and apoAII concentrations were measured by an immunoturbidimetric method with commercially available kits (Boehringer Mannheim) as previously described (18). The amount of apoAI residing in LpAI (hereafter termed LpAI-apoAI) was quantified using a differential electroimmunoassay with hydrated agarose gels containing monospecific antibodies against apoAI and apoAII (Sebia, Issy-Les-Moulineaux, France) as described (18). The amount of apoAI in LpAIAII (termed LpAIAII-apoAI) was then estimated by subtracting the concentration of LpAI from the immunoturbidimetrically measured total level of apoAI. The between-assay CVs for these assays were 3.5%, 2.1%, and 7.3% for apoAI, apoAII, and LpAI-apoAI, respectively.

LCAT activity was measured, using the endogenous plasma substrates, as the plasma cholesterol esterification rate (EST) (19). [3H]cholesterol complexed with albumin was added to plasma and plasma free cholesterol, and the labeled cholesterol was allowed to equilibrate overnight at 4°C. Subsequently, the labeled plasma was incubated for 1 h at 37°C as described (16). EST is expressed in absolute units of nanomoles per milliliter per hour. The between-assay CV was 8.3%.

Statistical methods.

All analyses were carried out using Stata 6 (StatCorp, College Station, TX). We examined the association between PLTP activity and various aspects of HDL using Spearman’s correlation coefficient (Table 1). To better quantify the magnitude of the difference in HDL parameters with a given difference in PLTP activity, we used multiple regression modeling with the HDL parameter as the outcome variable and PLTP activity (in units of 10 AU) as the explanatory variable. In these models, the regression coefficient (β) is the amount of difference in the HDL variable associated with 10 AU difference in PLTP activity (Table 2). We examined this association in all subjects, adjusting for diabetes, and also separately for subjects with and without diabetes. We tested whether the association between PLTP activity and HDL parameters differed in those with and without diabetes by including a diabetes-by-PLTP interaction term in these regression models, with diabetes as a dichotomous variable and PLTP as a continuous one. Additional adjustment was made for potential confounders by including them as covariates in these models. The same approach was used for examining the relationship between PLTP activity and apoB and apoB-containing lipoproteins.

Diabetic subjects had higher HDL cholesterol, apoAI, and LpAI-apoAI but not LpAIAII-apoAI levels (Table 3). Diabetes was associated with higher PLTP activity (mean 81 ± 15 AU in subjects without diabetes, 96 ± 19 AU in subjects with diabetes; P < 0.001), as reported previously (8).

The association of PLTP activity with aspects of HDL.

The Spearman correlation coefficients and regression coefficients for the association of PLTP activity with various aspects of HDL for each of the diabetes sex groups are shown in Tables 1 and 2, respectively. The most consistent findings across all four groups were that a higher PLTP activity was associated with lower levels of small HDL and higher levels of large HDL, apoAI, and apoAII.

As shown in Tables 1 and 2, not all of these associations were statistically significant within each of the four groups. In general, the difference in the HDL parameters for a 10-unit difference in PLTP activity was greatest in women with diabetes and lowest in women without diabetes. In subjects with diabetes, PLTP activity was also positively correlated with average HDL particle size. In women with diabetes, there was a positive association with HDL cholesterol.

Adjusting for other factors.

We examined whether the association between PLTP activity and HDL parameters might be attributable to any factors associated with PLTP activity other than PLTP activity itself. CET was positively correlated with PLTP activity (r = 0.16 in subjects without diabetes and r = 0.27 in subjects with diabetes; both P < 0.001). LCAT activity as measured by EST was also positively associated with PLTP activity in subjects both with and without diabetes (a difference of 4 units in EST with every 10-unit difference in PLTP activity in both groups; both P < 0.01). Other correlates of PLTP activity include CETP activity and LDL cholesterol levels (8). Adjusting for these variables, whether singly or simultaneously, increased rather than decreased the strength of the observed associations between PLTP activity and HDL parameters. For example, adjusting simultaneously for triglycerides, LDL cholesterol, CETP, CET, and EST increased the association between PLTP activity and large HDL (regression coefficient β = 0.86, P < 0.001 in subjects without diabetes; β = 1.4, P < 0.001 in subjects with diabetes on adjustment), reflecting that most of these positive correlates of PLTP activity are inversely correlated with HDL.

Adjustment for BMI and waist:hip ratio did not affect the associations observed either (data not shown). The strength of the relationship between PLTP activity and HDL parameters did not differ by triglyceride level (test for interaction between triglycerides and PLTP was >0.05 for all parameters). The conclusions were also unchanged when women who were on the pill or hormone replacement therapy were excluded from the analysis (data not shown) and when four subjects who were on β-blockers were excluded. Adjusting for alcohol consumption did not affect any of the associations between PLTP activity and HDL-related variables.

Role of elevated PLTP activity in differences in HDL between subjects with and without diabetes.

We examined whether the elevated PLTP activity in diabetes could explain differences in HDL in subjects with diabetes compared with subjects without diabetes. Adjusting for PLTP activity using multiple regression decreased the difference between subjects with and without diabetes in HDL size from 0.31 nm (9.41 nm vs. 9.1 nm, SD = 0.45) in subjects with diabetes to 0.26 nm, from 2.4 to 1.2 μmol/l for large HDL particle number, and from −3.2 to −2.2 μmol/l for small HDL particle number. The diabetic difference in apoAI was halved from 10.1 to 4.6 mg/dl. The diabetic difference in HDL cholesterol was reduced from 0.14 to 0.09 mmol/l.

PLTP activity and apoB.

PLTP activity was positively associated with apoB levels in subjects with diabetes (Spearman’s ρ = 0.28, P = 0.005 in men with diabetes; ρ = 0.22, P = 0.04 in women with diabetes) but was not associated with apoB in subjects without diabetes (ρ = 0.03, P = 0.8 in men; ρ = 0.1, P = 0.25 in women). Subjects with diabetes had a lower apoB than subjects without diabetes (9 mg/dl lower; P < 0.001). On adjustment for PLTP activity, the difference became more pronounced (12 mg/dl lower). We examined the relationship of PLTP activity to VLDL and LDL subclasses, total particle level, IDL level, and average VLDL and LDL particle size. The main findings were as follows: in subjects with diabetes, adjusted for sex, PLTP activity was positively associated with total VLDL particle number (P = 0.001), small VLDL (P = 0.001), IDL (P = 0.004), and total LDL particle number (P = 0.02) but not average VLDL or LDL particle size; in subjects without diabetes, PLTP activity was positively associated with small VLDL (P = 0.03).

In this large cross-sectional study, we demonstrated that higher PLTP activity is associated with having more large HDL and less small HDL and having more apoAI and apoAII. These associations were similar in direction in subjects with and without diabetes but were generally stronger in subjects with diabetes in whom PLTP activity was substantially elevated as previously reported (8). The observed associations were independent of other lipid and enzyme activities with which PLTP activity is correlated. The higher PLTP activity in patients with diabetes was sufficient to explain half of the diabetic-nondiabetic differences in apoAI and large HDL and less of the difference in HDL cholesterol, HDL size, and other aspects of HDL. Higher PLTP activity was also associated with higher apoB in subjects with diabetes.

The data are consistent with in vitro and in vivo studies that have demonstrated that PLTP has an effect on HDL remodeling (2–4,20–22). Our findings of a positive association between PLTP activity and large HDL and a negative association with small HDL are important because they demonstrate that in humans the role of PLTP in HDL metabolism is sufficiently critical to be detectable at the population level in a cross-sectional study. The large HDL that we measured by the NMR spectroscopy method is equivalent in size to the gradient gel electrophoresis subclasses 2b and 2a, and the small HDL that we defined is equivalent to 3b and 3c. PLTP co-elutes with large HDL in human serum (15), but there are few previous data on the association of PLTP activity with size-defined HDL subclasses in vivo. In a study of the effect of dietary change in 32 subjects, PLTP activity was positively associated with HDL2b and HDL2a subfractions and inversely correlated with HDL3a and HDL3b, consistent with our data on large and small HDL despite using different laboratory methods (23). In another study, HDL2b and HDL2a subfractions increased with increasing PLTP activity in 33 alcoholic patients in a cessation program (5).

Previous data on the relationship between PLTP activity and apoAI and apoAII are conflicting with PLTP activity being positively associated with apoAI but not apoAII in one study (24). In nonfasting samples in another study, PLTP activity was unrelated to apoAI but was positively correlated with apoAII (7). In our large fasting sample, we found a consistent pattern across all four diabetes sex groups of higher apoAI and apoAII with higher PLTP activity.

Previous data on the relationship of PLTP activity with HDL cholesterol and with LpAI and LpAIAII are also conflicting (7,24). We found PLTP activity to be positively correlated with LpAIAII-apoAI in subjects with diabetes also with LpAI-apoAI. The ratio of apoAI in LpAI to LpAIAII was unrelated to PLTP activity (data not shown). We found that PLTP activity was associated with HDL cholesterol in women with diabetes only. HDL cholesterol was positively associated with PLTP activity in fasting samples in some studies (5,6,24), but there was an inverse association between PLTP activity and HDL cholesterol in nonfasting samples in another (7). No association was found in two other studies (25,26). These inconsistent findings may reflect different assay methods, sample sizes, prevailing lipid levels, or the difference in fasting versus nonfasting assessment. In a sample of premenopausal women, PLTP activity was positively correlated with HDL2 and HDL3 cholesterol more strongly so in nonobese subjects (6).

Because our assay measures phospholipid transfer, our data are consistent with recent in vitro data demonstrating that the ability of PLTP to catalyze HDL conversion involves phospholipid transfer (27). HDL remodeling is a dynamic process, whereas we are assessing the relationship between HDL composition and PLTP activity at a single snapshot in time in the fasted state. Our study was cross-sectional and as such can only describe associations between PLTP activity and HDL. Demonstrating the role of PLTP in HDL remodeling definitively requires intervention studies. Other dynamic studies have supported the potential role for PLTP in HDL remodeling in humans by using infusions of insulin to reduce PLTP activity and observing a decrease in HDL cholesterol and apoAI (28,29). Our study design complements these intervention studies in that it reflects the prevailing levels of PLTP activity usually seen in the fasted state. However, as it is a cross-sectional study, it is of course possible that some other correlate of PLTP activity, rather than PLTP activity itself, is responsible for the associations observed. If so, then these correlates do not seem to be lipids, CETP activity, or LCAT activity. A possible correlate is hepatic lipase activity. However, PLTP activity and hepatic lipase activity have been reported to be unrelated, at least in subjects without diabetes, so this seems an unlikely explanation for our results (6). Lipoprotein lipase activity was found to be positively associated with PLTP activity in one study (6). In that study, PLTP activity was associated with HDL2 cholesterol when obese subjects were excluded and the association was not independent of lipoprotein lipase activity. We did not measure lipoprotein lipase activity as it is logistically difficult to do this on the large numbers of subjects as studied here, and so it cannot be excluded that it may be a partial explanation of the relationship of PLTP activity with HDL.

It is also the case that the relationship of lipoprotein composition to PLTP activity observed in this cross-sectional study may reflect a bidirectional association. For example, in pigs, the apoAI/apoAII molar ratio in HDL influences the PLTP-mediated remodeling (30). Unraveling the direction of associations found in cross-sectional studies requires additional approaches.

Our data also demonstrate that elevated PLTP activity in subjects with diabetes is sufficient to make an important contribution to their higher HDL particle size and apoAI levels, although it is also clear that elevated PLTP activity is not the sole explanation for this. Our data are cross-sectional and to demonstrate definitively that elevated PLTP activity is responsible for the altered HDL in patients with diabetes would require an intervention study altering PLTP activity and observing effects on HDL. The important therapeutic question is whether altering PLTP activity would have consequences for HDL metabolism that are related to atherosclerosis susceptibility. Plasma from transgenic mice overexpressing PLTP has enhanced ability to prevent accumulation of intracellular cholesterol in macrophages (4), but the generalizability of this to humans is unknown. Although increased generation of pre-β HDL would be expected to be antiatherogenic, it remains to be seen whether the composition of large HDL associated with elevated PLTP activity is such that it retains its antiatherogenic potential compared with large HDL at more usual levels of PLTP activity.

More recently, it has been suggested that increased PLTP might have atherogenic properties as there was reduced secretion of apoB-containing lipoproteins and atherosclerosis in PLTP-deficient mice (11). PLTP is probably involved in the lipidation of nascent apoB-containing particles within the secretory pathway. Our finding of a positive association between PLTP activity and apoB, total VLDL particle number, small VLDL, and IDL and LDL particle number is consistent with an important role for PLTP in apoB-containing lipoprotein secretion.

In conclusion, these data are consistent with an important role for PLTP in remodeling of HDL in the usual conditions pertaining in both the general population and subjects with diabetes. Elevated PLTP activity likely contributes to but is not the sole explanation for differences in HDL subclasses and apoAI in type 1 diabetes.

TABLE 1

Spearman correlation coefficients (ρ) for the association between PLTP activity and HDL-related variables

Nondiabetic men (ρ)Nondiabetic women (ρ)Diabetic men (ρ)Diabetic women (ρ)
HDL cholesterol (mmol/l) 0.13 −0.08 0.03 0.36* 
HDL size (nm) 0.10 −0.06 0.10 0.21† 
Small HDL (μmol/l) −0.10 −0.15 −0.10 −0.28‡ 
Intermediate HDL (μmol/l) 0.1 −0.03 0.01 0.01 
Large HDL (μmol/l) 0.21† 0.15 0.22† 0.40* 
ApoAI (mg/dl) 0.27† 0.0 0.11† 0.40* 
ApoAII (mg/dl) 0.24† 0.27‡ 0.13 0.34* 
LpAI (mg/dl) 0.11 −0.16 0.12 0.21 
LpAIAII (mg/dl) 0.3‡ 0.25‡ 0.10 0.36* 
Nondiabetic men (ρ)Nondiabetic women (ρ)Diabetic men (ρ)Diabetic women (ρ)
HDL cholesterol (mmol/l) 0.13 −0.08 0.03 0.36* 
HDL size (nm) 0.10 −0.06 0.10 0.21† 
Small HDL (μmol/l) −0.10 −0.15 −0.10 −0.28‡ 
Intermediate HDL (μmol/l) 0.1 −0.03 0.01 0.01 
Large HDL (μmol/l) 0.21† 0.15 0.22† 0.40* 
ApoAI (mg/dl) 0.27† 0.0 0.11† 0.40* 
ApoAII (mg/dl) 0.24† 0.27‡ 0.13 0.34* 
LpAI (mg/dl) 0.11 −0.16 0.12 0.21 
LpAIAII (mg/dl) 0.3‡ 0.25‡ 0.10 0.36* 

Data are means ± SE.

*

P < 0.001;

†

P < 0.05;

‡

P < 0.01.

View Large
TABLE 2

Regression coefficients (β; 95% CI) for the difference in HDL-related variables for every 10 AU higher PLTP activity adjusted for age

Nondiabetic men β (95% CI)Nondiabetic women β (95% CI)Nondiabetic men and women β (95% CI)Diabetic men β (95% CI)Diabetic women β (95% CI)Diabetic men and women β (95% CI)All subjects β (95% CI)
HDL cholesterol 0.05 (−0.01 to 0.1) −0.01 (−0.07 to 0.05) 0.02 (−0.02 to 0.06) 0.01 (−0.02 to 0.06) 0.07 (0.015–0.12)* 0.04 (0.01–0.08)* 0.025 (0.01–0.06)† 
HDL size (nm) 0.035 (−0.023 to 0.09) −0.01 (−0.79 to 0.57) 0.01 (−0.03 to 0.06) 0.05 (−0.01 to 0.95) 0.05 (−0.1 to 10.1) 0.048 (0.01–0.08)† 0.035 (0.01–0.06)* 
Small HDL (μmol/l) −0.57 (−1.5 to 0.43) −0.82 (−1.7 to 0.07) −0.65 (−1.3 to 0.01) −0.28 (−1 to 0.47) −1.0 (−1.65 to −0.34)† −0.64 (−1.1 to −0.1)* −0.67 (−0.1.1 to −0.27) 
Intermediate HDL (μmol/l) 0.42 (−0.28 to 1.13) −0.65 (−0.83 to 0.70) 0.19 (−0.32 to 0.7) −0.16 (−0.75 to 0.41) 0.05 (0.88–0.73) −0.05 (−0.5 to 0.4) 0.05 (−.29 to 0.38) 
Large HDL (μmol/l) 0.57 (−0.087 to 1.22) 0.47 (−0.15 to 1.1) 0.5 (0.07–0.96)* 0.64 (0.13–1.15)* 1.29 (0.72–1.9)‡ 0.97 (0.59–1.4)‡ 0.81 (0.52–1.1)‡ 
ApoAI (mg/dl) 3.8 (0.82–6.8)* 0.63 (−2.75 to 4.1) 2.3 (0.09–4.6)* 2.99 (0.5–5.4) 6.12 (3.1–9.1)‡ 4.6 (2.6–6.5)‡ 3.75 (2.3–5.2)‡ 
ApoAII (mg/dl) 0.90 (−0.2 to 2.0) 1.0 (−0.25 to 2.3) 0.97 (0.14–1.8)* 0.63 (−0.13 to 1.4) 1.6 (0.4–2.8)* 1.1 (0.4–1.8)‡ 1.1 (0.52–1.59)‡ 
LpAI (mg/dl) 1.0 (−0.94 to 3.0) −2.8 (−4.5 to 0.26) −0.4 (−1.9 to 1.13) 2.14 (0.46–3.8)* 2.3 (−0.1 to 4.6) 2.2 (0.76–3.6)† 1.2 (0.17–2.3)* 
LpAII (mg/dl) 2.8 (0.7–4.8)† 2.8 (0.9 to 4.7)† 2.7 (1.4–4.1)‡ 0.8 (−0.7 to 2.4) 3.9 (1.8–5.8)‡ 2.3 (1.1–3.7)‡ 2.7 (1.7–3.6)‡ 
Nondiabetic men β (95% CI)Nondiabetic women β (95% CI)Nondiabetic men and women β (95% CI)Diabetic men β (95% CI)Diabetic women β (95% CI)Diabetic men and women β (95% CI)All subjects β (95% CI)
HDL cholesterol 0.05 (−0.01 to 0.1) −0.01 (−0.07 to 0.05) 0.02 (−0.02 to 0.06) 0.01 (−0.02 to 0.06) 0.07 (0.015–0.12)* 0.04 (0.01–0.08)* 0.025 (0.01–0.06)† 
HDL size (nm) 0.035 (−0.023 to 0.09) −0.01 (−0.79 to 0.57) 0.01 (−0.03 to 0.06) 0.05 (−0.01 to 0.95) 0.05 (−0.1 to 10.1) 0.048 (0.01–0.08)† 0.035 (0.01–0.06)* 
Small HDL (μmol/l) −0.57 (−1.5 to 0.43) −0.82 (−1.7 to 0.07) −0.65 (−1.3 to 0.01) −0.28 (−1 to 0.47) −1.0 (−1.65 to −0.34)† −0.64 (−1.1 to −0.1)* −0.67 (−0.1.1 to −0.27) 
Intermediate HDL (μmol/l) 0.42 (−0.28 to 1.13) −0.65 (−0.83 to 0.70) 0.19 (−0.32 to 0.7) −0.16 (−0.75 to 0.41) 0.05 (0.88–0.73) −0.05 (−0.5 to 0.4) 0.05 (−.29 to 0.38) 
Large HDL (μmol/l) 0.57 (−0.087 to 1.22) 0.47 (−0.15 to 1.1) 0.5 (0.07–0.96)* 0.64 (0.13–1.15)* 1.29 (0.72–1.9)‡ 0.97 (0.59–1.4)‡ 0.81 (0.52–1.1)‡ 
ApoAI (mg/dl) 3.8 (0.82–6.8)* 0.63 (−2.75 to 4.1) 2.3 (0.09–4.6)* 2.99 (0.5–5.4) 6.12 (3.1–9.1)‡ 4.6 (2.6–6.5)‡ 3.75 (2.3–5.2)‡ 
ApoAII (mg/dl) 0.90 (−0.2 to 2.0) 1.0 (−0.25 to 2.3) 0.97 (0.14–1.8)* 0.63 (−0.13 to 1.4) 1.6 (0.4–2.8)* 1.1 (0.4–1.8)‡ 1.1 (0.52–1.59)‡ 
LpAI (mg/dl) 1.0 (−0.94 to 3.0) −2.8 (−4.5 to 0.26) −0.4 (−1.9 to 1.13) 2.14 (0.46–3.8)* 2.3 (−0.1 to 4.6) 2.2 (0.76–3.6)† 1.2 (0.17–2.3)* 
LpAII (mg/dl) 2.8 (0.7–4.8)† 2.8 (0.9 to 4.7)† 2.7 (1.4–4.1)‡ 0.8 (−0.7 to 2.4) 3.9 (1.8–5.8)‡ 2.3 (1.1–3.7)‡ 2.7 (1.7–3.6)‡ 
*

P < 0.05;

†

P < 0.01;

‡

P < 0.001 for the association between PLTP activity and lipoprotein parameters adjusted for age using multiple linear regression and also adjusted for sex and diabetes where appropriate.

View Large
TABLE 3

Subject characteristics by diabetes and sex

Men
Women
Men and women
Nondiabetic (n = 89; mean [SD])Diabetic (n = 101; mean [SD])P value*Nondiabetic (n = 105; mean [SD])Diabetic (n = 94; mean [SD])P valueNondiabetic (n = 194; mean [SD])Diabetic (n = 195; mean [SD])P value†
Age (years) 37.8 (4) 38.0 (4) 0.7 37.9 (4) 37.5 (4) 0.5 38 (0.3) 38 (0.3) 0.8 
BMI (kg/m2) 24.9 (3.4) 25.4 (3.2) 0.3 25.6 (5.6) 25.3 (3.8) 0.6 25.3 (4.7) 25.3 (3.5) 0.9 
HbAlc (%) 5.3 (0.4) 8.4 (1.2) < 0.001 5.3 (0.4) 9.1 (1.7) < 0.001 5.3 (0.4) 8.9 (1.5) < 0.001 
Diabetes duration (years) — 23 (7.6) — — 24 (7.4) — — 23 (7.5) — 
Triglyceride (mmol/l)‡ 1.32 (1.0–1.7) 1.03 (0.9–1.5) 0.04 0.96 (0.7–1.3) 0.97 (0.7–1.2) 0.9 1.11 (0.8–1.5) 1.06 (0.8–1.4) 0.14 
LDL cholesterol (mmol/l) 3.3 (1.1) 3.0 (0.96) 0.03 2.9 (0.83) 2.8 (0.87) 0.3 3.1 (0.96) 2.9 (0.92) 0.03 
ApoB (mg/dl) 99 (28) 84 (24) < 0.001 82 (20) 78 (22) 0.2 90 (26) 81 (24) < 0.001 
HDL cholesterol (mmol/l) 1.6 (0.34) 1.7 (0.34) 0.01 1.8 (0.39) 2.0 (0.49) 0.04 1.7 (0.41) 1.8 (0.46) 0.001 
ApoAI (mg/dl) 143 (22) 152 (23) 0.01 153 (23) 165 (24) 0.002 149 (23) 158 (27) < 0.001 
ApoAII (mg/dl) 42 (8) 40 (7) 0.2 39 (9) 40 (11) 0.6 40 (8) 40 (9) 0.2 
LpAI (mg/dl) 51 (14) 60 (16) < 0.001 66 (16) 76 (22) < 0.001 59 (17) 67 (20) < 0.001 
LpAIAII (mg/dl) 92.7 (15) 92.1 (14) 0.7 87.6 (13) 89.1 (20) 0.5 90 (14) 90.6 (17) 0.8 
% on pill, implant, or HRT (SE) — — — 24 (4) 18 (4) 0.3 — —  
% obese (SE) 7 (3) 8 (3) 0.8 21 (4) 8 (3) 0.02 14 (2) 8 (2) 0.08 
% with albuminuria (SE) 3 (2) 21 (4) 0.001 3 (2) 9 (3) 0.8 3 (1) 16 (3) < 0.001 
% drinking >21 units/week (SE) 38 (5) 25 (4) 0.04 7 (2) 11 (3) 0.3 21 (3) 18 (3) 0.2 
Men
Women
Men and women
Nondiabetic (n = 89; mean [SD])Diabetic (n = 101; mean [SD])P value*Nondiabetic (n = 105; mean [SD])Diabetic (n = 94; mean [SD])P valueNondiabetic (n = 194; mean [SD])Diabetic (n = 195; mean [SD])P value†
Age (years) 37.8 (4) 38.0 (4) 0.7 37.9 (4) 37.5 (4) 0.5 38 (0.3) 38 (0.3) 0.8 
BMI (kg/m2) 24.9 (3.4) 25.4 (3.2) 0.3 25.6 (5.6) 25.3 (3.8) 0.6 25.3 (4.7) 25.3 (3.5) 0.9 
HbAlc (%) 5.3 (0.4) 8.4 (1.2) < 0.001 5.3 (0.4) 9.1 (1.7) < 0.001 5.3 (0.4) 8.9 (1.5) < 0.001 
Diabetes duration (years) — 23 (7.6) — — 24 (7.4) — — 23 (7.5) — 
Triglyceride (mmol/l)‡ 1.32 (1.0–1.7) 1.03 (0.9–1.5) 0.04 0.96 (0.7–1.3) 0.97 (0.7–1.2) 0.9 1.11 (0.8–1.5) 1.06 (0.8–1.4) 0.14 
LDL cholesterol (mmol/l) 3.3 (1.1) 3.0 (0.96) 0.03 2.9 (0.83) 2.8 (0.87) 0.3 3.1 (0.96) 2.9 (0.92) 0.03 
ApoB (mg/dl) 99 (28) 84 (24) < 0.001 82 (20) 78 (22) 0.2 90 (26) 81 (24) < 0.001 
HDL cholesterol (mmol/l) 1.6 (0.34) 1.7 (0.34) 0.01 1.8 (0.39) 2.0 (0.49) 0.04 1.7 (0.41) 1.8 (0.46) 0.001 
ApoAI (mg/dl) 143 (22) 152 (23) 0.01 153 (23) 165 (24) 0.002 149 (23) 158 (27) < 0.001 
ApoAII (mg/dl) 42 (8) 40 (7) 0.2 39 (9) 40 (11) 0.6 40 (8) 40 (9) 0.2 
LpAI (mg/dl) 51 (14) 60 (16) < 0.001 66 (16) 76 (22) < 0.001 59 (17) 67 (20) < 0.001 
LpAIAII (mg/dl) 92.7 (15) 92.1 (14) 0.7 87.6 (13) 89.1 (20) 0.5 90 (14) 90.6 (17) 0.8 
% on pill, implant, or HRT (SE) — — — 24 (4) 18 (4) 0.3 — —  
% obese (SE) 7 (3) 8 (3) 0.8 21 (4) 8 (3) 0.02 14 (2) 8 (2) 0.08 
% with albuminuria (SE) 3 (2) 21 (4) 0.001 3 (2) 9 (3) 0.8 3 (1) 16 (3) < 0.001 
% drinking >21 units/week (SE) 38 (5) 25 (4) 0.04 7 (2) 11 (3) 0.3 21 (3) 18 (3) 0.2 
*

P value for the difference between diabetic and nondiabetic subjects within each sex, adjusted for age using multiple linear regression;

†

P value also adjusted for sex;

‡

median and interquartile range is shown as data skewed; HRT, hormone replacement therapy. Some data are reprinted with permission from the American College of Cardiology (J Am Coll Cardiol 36:2160–2167, 2000).

View Large

This work was funded by a project grant from the British Heart Foundation. We thank T. van Gent and L.M. Scheek for expert assistance with the CETP and EST assays, respectively.

1.
Rye KA, Clay MA, Barter PJ: Remodelling of high density lipoproteins by plasma factors.
Atherosclerosis
 
145
:
227
–238,
1999
2.
Jauhiainen M, Metso J, Pahlman R, Blomqvist S, van Tol A, Ehnholm C: Human plasma phospholipid transfer protein causes high density lipoprotein conversion.
J Biol Chem
 
268
:
4032
–4036,
1993
3.
Foger B, Santamarina Fojo S, Shamburek RD, Parrot CL, Talley GD, Brewer HB Jr: Plasma phospholipid transfer protein: adenovirus-mediated overexpression in mice leads to decreased plasma high density lipoprotein (HDL) and enhanced hepatic uptake of phospholipids and cholesteryl esters from HDL.
J Biol Chem
 
272
:
27393
–27400,
1997
4.
van Haperen R, van Tol A, Vermeulen P, Jauhiainen M, van Gent T, van den BP, Ehnholm S, Grosveld F, van der KA, de Crom R: Human plasma phospholipid transfer protein increases the antiatherogenic potential of high density lipoproteins in transgenic mice.
Arterioscler Thromb Vasc Biol
 
20
:
1082
–1088,
2000
5.
Lagrost L, Athias A, Herbeth B, Guyard-Dangremont V, Artur Y, Paille F, Gambert P, Lallemant C: Opposite effects of cholesteryl ester transfer protein and phospholipid transfer protein on the size distribution of plasma high density lipoproteins: physiological relevance in alcoholic patients.
J Biol Chem
 
271
:
19058
–19065,
1996
6.
Murdoch SJ, Carr MC, Hokanson JE, Brunzell JD, Albers JJ: PLTP activity in premenopausal women: relationship with lipoprotein lipase, HDL, LDL, body fat, and insulin resistance.
J Lipid Res
 
41
:
237
–244,
2000
7.
Tahvanainen E, Jauhiainen M, Funke H, Vartiainen E, Sundvall J, Ehnholm C: Serum phospholipid transfer protein activity and genetic variation of the PLTP gene.
Atherosclerosis
 
146
:
107
–115,
1999
8.
Colhoun HM, Scheek LM, Rubens MR, Van Gent T, Underwood SR, Fuller JH, Van Tol A: Lipid transfer protein activities in type 1 diabetic patients without renal failure and non-diabetic controls and their association with coronary artery calcification.
Diabetes
 
50
:
652
–659,
2001
9.
Kahri J, Groop PH, Viberti G, Elliott T, Taskinen MR: Regulation of apolipoprotein A-I-containing lipoproteins in IDDM.
Diabetes
 
42
:
1281
–1288,
1993
10.
Colhoun HM, Otvos JD, Rubens MB, Taskinen M-R, Underwood SR, Fuller JH: Lipoprotein subclasses and particle sizes and their relationship with coronary artery calcification in men and women with and without type 1 diabetes.
Diabetes
 
51
:
1949
–1956,
2002
11.
Jiang XC, Qin S, Qiao C, Kawano K, Lin M, Skold A, Xiao X, Tall AR: Apolipoprotein B secretion and atherosclerosis are decreased in mice with phospholipid-transfer protein deficiency.
Nat Med
 
7
:
847
–852,
2001
12.
Allain CC, Poon LS, Chan CS, Richmond W, Fu PC: Enzymatic determination of total serum cholesterol.
Clin Chem
 
20
:
470
–475,
1974
13.
Fossati P, Prencipe L: Serum triglycerides determined colorimetrically with an enzyme that produces hydrogen peroxide.
Clin Chem
 
28
:
2077
–2080,
1982
14.
Sugiuchi H, Uji Y, Okabe H, Irie T, Uekama K, Kayahara N, Miyauchi K: Direct measurement of high-density lipoprotein cholesterol in serum with polyethylene glycol-modified enzymes and sulfated alpha-cyclodextrin.
Clin Chem
 
41
:
717
–723,
1995
15.
Speijer H, Groener JE, van Ramshorst E, van Tol A: Different locations of cholesteryl ester transfer protein and phospholipid transfer protein activities in plasma.
Atherosclerosis
 
90
:
159
–168,
1991
16.
Dullaart RP, Riemens SC, Scheek LM, van Tol A: Insulin decreases plasma cholesteryl ester transfer but not cholesterol esterification in healthy subjects as well as in normotriglyceridaemic patients with type 2 diabetes.
Eur J Clin Invest
 
29
:
663
–671,
1999
17.
Otvos JD: Measurement of lipoprotein subclass profiles by nuclear magnetic resonance spectroscopy. In
Handbook of Lipoprotein Testing
. Rifai N, Warnick GR, Dominiczak MH, Eds. Washington, DC, AACC Press,
2000
, p.
609
–223
18.
Tilly-Kiesi M, Kahri J, Pyorala T, Puolakka J, Luotola H, Lappi M, Lahdenpera S, Taskinen MR: Responses of HDL subclasses, Lp(A-I) and Lp(A-I:A-II) levels and lipolytic enzyme activities to continuous oral estrogen-progestin and transdermal estrogen with cyclic progestin regimens in postmenopausal women.
Atherosclerosis
 
129
:
249
–259,
1997
19.
Channon KM, Clegg RJ, Bhatnagar D, Ishola M, Arrol S, Durrington PN: Investigation of lipid transfer in human serum leading to the development of an isotopic method for the determination of endogenous cholesterol esterification and transfer.
Atherosclerosis
 
80
:
217
–226,
1990
20.
Settasatian N, Duong M, Curtiss LK, Ehnholm C, Jauhiainen M, Huuskonen J, Rye KA: The mechanism of the remodeling of high density lipoproteins by phospholipid transfer protein.
J Biol Chem
 
276
:
26898
–26905,
2001
21.
Jiang XC, Bruce C, Mar J, Lin M, Ji Y, Francone OL, Tall AR: Targeted mutation of plasma phospholipid transfer protein gene markedly reduces high-density lipoprotein levels.
J Clin Invest
 
103
:
907
–914,
1999
22.
Ehnholm S, van Dijk KW, van ’t HB, van der ZA, Olkkonen VM, Jauhiainen M, Hofker M, Havekes L, Ehnholm C: Adenovirus mediated overexpression of human phospholipid transfer protein alters plasma HDL levels in mice.
J Lipid Res
 
39
:
1248
–1253,
1998
23.
Lagrost L, Mensink RP, Guyard-Dangremont V, Temme EH, Desrumaux C, Athias A, Hornstra G, Gambert P: Variations in serum cholesteryl ester transfer and phospholipid transfer activities in healthy women and men consuming diets enriched in lauric, palmitic or oleic acids.
Atherosclerosis
 
142
:
395
–402,
1999
24.
Cheung MC, Wolfbauer G, Brown BG, Albers JJ: Relationship between plasma phospholipid transfer protein activity and HDL subclasses among patients with low HDL and cardiovascular disease.
Atherosclerosis
 
142
:
201
–205,
1999
25.
Mero N, van Tol A, Scheek LM, van Gent T, Labeur C, Rosseneu M, Taskinen MR: Decreased postprandial high density lipoprotein cholesterol and apolipoproteins A-I and E in normolipidemic smoking men: relations with lipid transfer proteins and LCAT activities.
J Lipid Res
 
39
:
1493
–1502,
1998
26.
Desrumaux C, Athias A, Bessede G, Verges B, Farnier M, Persegol L, Gambert P, Lagrost L: Mass concentration of plasma phospholipid transfer protein in normolipidemic, type IIa hyperlipidemic, type IIb hyperlipidemic, and non-insulin-dependent diabetic subjects as measured by a specific ELISA.
Arterioscler Thromb Vasc Biol
 
19
:
266
–275,
1999
27.
Huuskonen J, Olkkonen VM, Ehnholm C, Metso J, Julkunen I, Jauhiainen M: Phospholipid transfer is a prerequisite for PLTP-mediated HDL conversion.
Biochemistry
 
39
:
16092
–16098,
2000
28.
van Tol A, Ligtenberg JJ, Riemens SC, van Haeften TW, Reitsma WD, Dullaart RP: Lowering of plasma phospholipid transfer protein activity by acute hyperglycaemia-induced hyperinsulinaemia in healthy men.
Scand J Clin Lab Invest
 
57
:
147
–157,
1997
29.
Riemens SC, van Tol A, Sluiter WJ, Dullaart RPF: Plasma phospholipid transfer protein activity is related to insulin resistance: impaired acute lowering by insulin in obese NIDDM patients.
Diabetologia
 
41
:
929
–934,
1998
30.
Pussinen PJ, Jauhiainen M, Ehnholm C: ApoA-II/apoA-I molar ratio in the HDL particle influences phospholipid transfer protein-mediated HDL interconversion.
J Lipid Res
 
38
:
12
–21,
1997

Address correspondence and reprint requests to Helen M. Colhoun, EURODIAB, Department of Epidemiology and Public Health, University College London Medical School, 1-19 Torrington Place, London WC1E 6BT. E-mail: helen@public-health.ucl.ac.uk.

Received for publication 28 February 2002 and accepted in revised form 19 July 2002.

J.D.O. is an employee of, a member of the Board of Directors for, and holds stock in Liposcience.

apo, apolipoprotein; AU, arbitrary units; CET, cholesterol ester transfer; CETP, cholesterol ester transfer protein; CV, coefficient of variation; EST, plasma cholesterol esterification rate; IDL, intermediate-density lipoprotein; LCAT, lecithin:cholesterol acyl transferase; NMR, nuclear magnetic resonance; PLTP, phospholipid transfer protein.

DIABETES
2002