Although insulin resistance has been suggested to be a basic defect, little is known ... investigated the relation between insulin sensitivity, lipoprotein distribution, ...
E p i d e m i o I o g y / H e a 11 h N A L
Se rv ice s/ Psy ch o so cia I
R e s e a r c h
A R T I C L E
LDL Size Distribution in Relation to Insulin Sensitivity and Lipoprotein Pattern in Young and Healthy Subjects ANDREAS AMBROSCH, MD ISABEL MOHLEN, MD DANIEL KOPF, MD WOLFGANG AUGUSTUS, PHD
JUTTA DlERKES, PHD WOLFGANG KONIG, MD CLAUS LULEY, MD HENDRIK LEHNERT, MD
OBJECTIVE — Smaller LDL particles are associated with an increased risk for coronary artery disease and have been found predominantly in subjects with the insulin resistance syndrome. Although insulin resistance has been suggested to be a basic defect, little is known about the relation between this predisposing factor (and associated metabolic disturbances) and LDL size distribution in young and metabolically healthy subjects. In the present study, we investigated the relation between insulin sensitivity, lipoprotein distribution, and LDL patterns in young adults to increase the understanding of the development of metabolic risk factors in an early phase of the life span. RESEARCH DESIGN A N D METHODS— Young, clinically healthy subjects (n = 50; age 21.1-30.6 years) were enrolled in the study. Glucose metabolism was characterized by peripheral insulin sensitivity assessed by a hyperinsulinemic-euglycemic clamp and by levels of fasting insulin, C-peptide, and glucose. Lipoproteins were measured, and LDL fractions were additionally characterized by the diameter of the major LDL peak, estimated by 2-16% polyacrylamide gradient gel electrophoresis. Cholesterol ester transfer was estimated with a fluorescent spectroscopic method that measures the transfer of fluorescent cholesteryl linoleate between exogenous donor and acceptor particles. In this assay system, cholesterylester transfer protein (CETP) activity was only influenced by the plasma CETP concentration therefore reflecting more likely the CETP mass. RESULTS — In the entire study group, 47 subjects had LDL phenotype A (LDL diameter >25.75 nm) and 3 subjects had an intermediate phenotype (25.50-25.75 nm). An interrelation between LDL size and LDL triglyceride (LDL-TG) per apolipoprotein (apo) B (Spearman's rank correlation analysis; r = —0.78; P < 0.001) or LDL cholesterol ester (CE) per apoB (r = 0.58, P < 0.001) was found, and 39% of the plasma samples studied were characterized by a monodispersed LDL pattern. Furthermore, LDL diameters correlated negatively with total TGs (men: r = —0.52, P < 0.001; women: r = —0.61, P < 0.001) and positively with insulin sensitivity (total population: r = 0.54, P < 0.001). In addition, LDL size was inversely related to the [VLDL + LDL cholesterol (CH)]/HDL-CH ratio and positively to the HDL-CE/TG ratio, which were both related vice versa to CETP activity levels. A direct relation between CETP activity levels and LDL size or composition was not observed. In a linear regression analysis including parameters of lipoprotein metabolism (TG, HDL cholesterol, CETP activity level), glucose metabolism (insulin sensitivity, fasting insulin), and sex, only TGs predicted significantly for 62% of LDL size variability. If the total study population was evaluated according to quintiles of insulin sensitivity, increasing TGs (analysis of variance, Scheffe test; P < 0.05) and
From the Institutes of Clinical Chemistry (A.A., J.D., C.L.) and Microbiology (A.A., WK.), and the Division of Endocrinology (I.M., D.K., H.L.), the Departments of Pathobiochemistry and Medicine (WA.), University Hospital of Magdeburg, Magdeburg, Germany. Address correspondence and reprint requests to Andreas Ambrosch, MD, Institute of Microbiology, University Hospital of Magdeburg, Leipziger Str. 44, 39120 Magdeburg, Germany. E-mail: andreas.ambrosch® medizin.uni-magdeburg.de. Received for publication 9 February 1998 and accepted in revised form 13 August 1998. Abbreviations: ANOVA, analysis of variance; apo, apolipoprotein; CE, esterified cholesterol; CETP, cholesterylester transfer protein; FC, unesterified cholesterol; FFA, free fatty acid; NBD, nitrobenzooxadiazol-fluorophor; PAGGE, nondenaturating polyacrylamide gradient gel electrophoresis; TG, triglyceride. A table elsewhere in this issue shows conventional and Systeme International (SI) units and conversion factors for many substances.
DIABETES CARE, VOLUME 21, NUMBER 12, DECEMBER
1998
CETP activitiy levels (P < 0.05) were combined with decreasing LDL particle diameters (P < 0.05) and with a preponderance of a monodispersed LDL pattern (60%) in the most insulin-resistant group. CONCLUSIONS — Among parameters of the lipoprotein and glucose metabolism, total TG is the single most important factor affecting LDL size variability, even in young adults. If the study population is evaluated according to insulin sensitivity, lipoprotein pattern is altered in a more atherogenic manner in the most insulin-resistant subjects. In this group, increasing TG and CETP activity levels are associated with decreasing LDL particle diameters and preponderance of a monodispersed LDL pattern. Although increasing CETP levels are combined with this particular lipoprotein profile, a direct relation to LDL size and composition is not found. Diabetes Care 21:2077-2084,1998
A
nalysis of LDL particle size distributions by nondenaturating gradient gel electrophoresis (PAGGE) has identified multiple, distinct LDL subclasses and has shown that LDL cholesterol in most individuals can be characterized by a predominance of larger LDL or smaller LDL cholesterol. More exactly, Austin et al. (1) described two distinct LDL subclass phenotypes based on the LDL diameter: pattern A with a major diameter of >25.5 nm and pattern B with a diameter ^ 2 5 . 5 nm (small, dense LDL). Epidemiological studies suggested that a predominance of LDL pattern B is highly atherogenic and may become a nontraditional risk factor of coronary artery disease (2-4). Factors regulating individual LDL size are not completely understood. Genetic influences on LDL particle size are supported by data from complex segregation analysis in families, hereditability analysis in twins, and linkage analysis (4-6). Substantial data suggesting that nongenetic factors modify the expression of the phenotype have recently emerged. Several studies have shown that increased plasma triglyceride (TG) levels and decreased HDL cholesterol are strongly related to the prevalence of small, dense LDL cholesterol. In addition, a higher prevalence
2077
Insulin sensitivity and lipoprotein pattern
of smaller LDL particles has been reported in type 2 diabetes, suggesting a role of insulin resistance in LDL size reduction (7-9). Further factors contributing to the LDL size distribution may be the intravascular lipolytic system (10,11), the rates of catabolism of LDL cholesterol by the LDL receptor-dependent pathway (12), esterification rates in HDL cholesterol (13), and cholesterol ester transfer activity from LDL to VLDL cholesterol mediated by cholesterylester transfer protein (CETP) (13,14). The aim of the present study was to investigate the relation between insulin sensitivity assessed by hyperinsulinemic-euglycemic clamp, parameters of the lipoprotein metabolism, and LDL components. In this context, we analyzed LDL composition as well as major particle diameter and the mono- or polydispersity of the lipoprotein pattern within the LDL range. Since cholesterol ester transfer between lipoproteins plays a crucial role in intravascular LDL remodeling, the CETP-dependent transfer in relation to LDL lipoprotein composition and size has been evaluated. Young and healthy adults were used as study subjects because the effects of medication and chronic diseases on lipoprotein and glucose metabolism did not need to be considered.
RESEARCH DESIGN AND METHODS Study subjects A total of 50 subjects (24 men, 26 women) were enrolled in the study. Subjects were considered eligible if they were between ages 18 and 32 years, were clinically healthy, and did not use any medication besides oral contraceptives. Information about smoking habits and family history of diabetes was obtained using a standardized questionnaire. All subjects were students recruited from the University of Magdeburg. All participants gave their written informed consent, and the ethics committee of the University of Magdeburg approved the study. Determination of insulin sensitivity A euglycemic-hyperinsulinemic glucose clamp was performed for each subject. The subjects were instructed to choose carbohydrate-rich diets for 3 days before each test. Women were studied during the first 10 days of their menstrual cycle to minimize hormonal influence on insulin sensitivity at different times of the cycle. Systolic and diastolic blood pressure was measured twice after at least 10 min in the supine 2078
position with a random zero sphygmomanometer. Insulin sensitivity was determined with the euglycemic glucose clamp technique (15). Briefly, the subjects were admitted to the Department of Medicine, Clinic of Endocrinology, at 8:00 A.M. after an overnight fast of 12 h. A plastic 18-gauge French cannula was inserted into a left antecubital vein for infusions, and a 21-gauge French cannula was inserted retrogradely into a dorsal vein of the right hand. Basal blood samples were then collected for determination of plasma glucose, C-peptide, insulin, lipid profile, and CETP activity. After a 1-h rest to reestablish basal conditions, regular human insulin (Human S, Lilly, Bad Homburg, Germany) was administered at 1.0 mU • kg"1 • min"1 for 2 h. Capillary blood samples were obtained at 5min intervals for estimation of the plasma glucose. The samples were centrifuged immediately and analyzed with the glucose oxidase method in a Beckman glucose analyzer (Beckman, Munich). Plasma glucose was maintained at the fasting level by a variable exogenous infusion of 20% glucose solution. Serum samples for determination of insulin, C-peptide, lipoprotein profile, and CETP activity were separated as soon as clotting was complete and were stored at —20°C until analysis. Insulin sensitivity was calculated from the glucose utilization rate per kilogram body weight (micromole per kilogram per minute) during the hyperinsulinemic steady state.
mined by measurement of cholesterol, TG, unesterified polyester (FC), and apolipoproteins on an automated Hitachi 911 analyzer (Boehringer Mannheim, Mannheim, Germany). The LDL composition was calculated from the bottom fractions. Lipoprotein component assays. Total apolipoproteins (apo) AI, B, CIII, and E were analyzed by immunoturbidimetric assays using goat antihuman polyclonal antisera (Greiner, Bad Homburg, Germany). We used the CHOD-PAP (cholesterol-esteraseperoxidase-phenolaminoantipyrin) method (Choi kit, Boehringer) for measurement of cholesterol, the GPO-PAP (glycerolphosphate-oxidase-phenolaminoantipyrin) method for TGs and free glycerol (Triglyceride GPO-PAP kit, Boehringer), the CODPAP (cholesterol-oxidase-phenolaminoantipyrin) method (free cholesterol C kit, WAK, Bad Homburg, Germany) for free cholesterol, and the ACS-ACOD (acyl-CoA-synthase-acy-CoA-oxidase) method (NEFA C, WAK) for measurement of free fatty acids (FFAs). The esterified cholesterol (CE) content was calculated from total cholesterol and total FC concentrations. PAGGE. For estimation of LDL particle size from the density =1.21 supranatant, 2-16% PAGGE was performed in a modification of a published protocol (16). Briefly, the stock solutions, which were made up in TE buffer (Tris 210 mmol/1, sodium azide 3 mmol/1, disodium EDTA 3 mmol/1 [pH 8.35]), included the following: the high limit solution acrylamide 152.3 g/1, bis-acrylamide 7.7 g/1 (16% total, 4% crosslinker), and glycerol 260 g/1 (all from Sigma, Munich), and Analytical methods Insulin and C-peptide. C-peptide and the low limit solution acrylamide, 16 g/1 insulin were determined by using commer- and bis-acrylamide 4 g/1 (2% total, 4% cially available radioimmunoassays (CIS crosslinker). The high and the low limit Biointernational, Paris). The cross-reactivity solutions were made immediately before the of the insulin assay with proinsulin was gradient was cast and contained 1 vol high 200 mg/dl) 169 (130-234) Total cholesterol (mg/dl) Apolipoprotein 66 (47-104) B (mg/dl) 11.3(8.2-25.4) c m (g/i) 3.4(1.8-6.4) E(g/1) AI (mg/dl) 135 (102-209) 50(11-107) VLDL TG (mg/dl) VLDL cholesterol (mg/dl) 14 (3-25) LDL cholesterol (mg/dl) 94 (63-154) LDL-TG (mg/dl) 21(6-40) 70(44.118) LDL-CE (mg/dl) LDL-FC (mg/dl) 26 (18-38) LDL-apoB (mg/dl) 58 (45-91) LDL cholesterol/apoB 1.59 (0.94-2.08) LDL-TG/apoB 0.60(0.12-0.76) LDL-CE/apoB 1.61 (0.82-1.66) LDL-FC/apoB 0.41 (0.29-0.65) LDL size (nm) 26.74 (25.84-28.20) HDL cholesterol (mg/dl) 52(41-91) Low HDL cholesterol (%) 4 (
