Findings

The MetS is characterized by three or more of the following features: increased waist circumference (WC), high triglycerides (TGs), lowered high density cholesterol (HDL-C), hypertension and impaired fasting glucose [1]. Shifts in overall lipoprotein diameter to smaller LDL and HDL particles, and to larger VLDL particles have additionally been shown to associate with the individual features of the MetS [2].

Our previous research grouped individuals according to the fasting diameters of their VLDL, LDL and HDL particles. Two groups emerged where individuals, on average, met the adult treatment panel (ATP) III criteria for the MetS. Both groups had small LDL diameters (mean = 19.9 nm), but one group with larger VLDL diameters (65.74 vs 53.1 nm) showed higher IR as measured by homeostatic model assessment (HOMA-IR; Table 1); and increased features of the MetS associated with higher IR: higher fasting glucose, TG, and WC (Table 1).

Table 1 Lipid, demographic, genetic and MetS-related characteristics for participants by group
Full size table

LPL hydrolyzes TGs from VLDL forming LDL, thus is a source of lipoprotein size heterogeneity. We therefore aimed to examine whether there were differences in two variants in the LPL gene between two groups of MetS individuals, defined by their fasting lipoprotein diameter pattern.

Materials and Methods

Participants and study design

The study sample includes 251 men and women participating in the general population of the GOLDN study. The majority of participants were re-recruited from three-generational pedigrees from two field centers in Minnesota and Utah. Nearly all individuals were of European descent. The study protocol has been published elsewhere [3], and was approved by the institutional review boards of the Universities of Alabama, Minnesota, and Utah and Tufts University.

Latent class analysis (LCA) was used to group all GOLDN individuals according to the diameter of their fasting (8-hour fast) VLDL, LDL and HDL particles, determined by nuclear resonance spectroscopy. LCA is a form of cluster analysis that groups individuals based on similarities within specified measures, and uses an iterative approach to determine the number of groups (classes) that best explains the population variance and covariance in the measures [4]. Initially, all participants are considered to be in a single class and classes are added until an additional class does not improve the fit of the model, as assessed using the log Bayes Factor (logBF). The LCA was implemented using the TRAJ Procedure in SAS [5]. As the LCA uses maximum likelihood estimation, the absolute diameter was standardized within each fraction of HDL, LDL and VLDL to ease optimization.

LCA grouped all GOLDN individuals into 8 classes (groups) according their fasting VLDL, LDL and HDL diameters. For two groups (N = 251), the average number of MetS features per member (3.5; Table 1) indicated that the majority of individuals within these groups met ATP-III criteria for the MetS. The two groups were not distinguished by the diameter of their LDL or HDL particles but by that of their VLDL particles (Table 1). The current analysis examined these groups for differences in three variants within the LPL gene.

Genotype data

Genomic data was isolated from blood samples using routine DNA isolation sets (Qiagen, Valencia, CA, USA), using a TaqMan assay with allele-specific probes on the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA, USA) according to the standardized laboratory protocols.

SNP selection

Three nonsynonymous cSNPs at LPL codons 9, 291, and 447, namely, D9N, M107 and S447X were chosen for literature reports of genetic associations or lipid-related biological function. The minor allele frequency (MAF) of M107 was 0, and so excluded from analysis.

Statistical analysis

All analyses were conducted in SAS v9.2. Hardy-Weinberg equilibrium (HWE) was determined using χ2 analysis. For genotype-phenotype associations, regression analysis was conducted using the "PROC GENMOD" command with genotype frequency, age, age2 and age3, sex, and center of data collection as predictors, and LCA group as the outcome. Pedigree was treated as a fixed random effect. To increase power, carriers of the minor allele for each genotype were compared to non carriers.

Results

Participant characteristics

Lipoprotein, demographic and MetS-related characteristics are shown in Table 1.

LPL genotypes

Genotype frequencies did not deviate from HWE expectations (P < .05). At the D9N locus, individuals were either homozygous for D9 allele (frequency = .94) or heterozygous (frequency = .06). The S447X locus showed individuals who were homozygous for the major allele (S447; frequency = .84), the minor allele (X447; frequency = .08) or heterozygous (frequency = .16).

Genotype-phenotype associations

The D9N minor allele (MAF = .10) did not show a group association (P = .36; Table 1); however in the fully adjusted model (controlling for age, sex, data collection center and pedigree) there were significantly more carriers of the minor allele 447X variant (MAF = .08) in group 2, than group 1 (P = .04).

Discussion

This study provides evidence that the minor allele of the LPL S774X variant is associated with larger VLDL diameters within groups defined by a lipoprotein diameter pattern which includes small LDL and HDL particles. The majority of individuals within these groups met ATP-III criteria for the MetS, and the pattern of large VLDL in addition to the small LDL and HDL particles additionally associates with higher IR, as well as the highest WC, fasting glucose and TGs, and diabetes prevalence. We did not find any association with lipoprotein diameter pattern and the D9N polymorphism.

The LPL S447X variant has been implicated in functional relevance to lipid parameters [6]. The main functions of the LPL S447X variant are the hydrolysis of plasma TGs, and in mediating the clearance of atherogenic remnant lipoproteins from the circulation [7]. Total missense, and nonsense mutations lead to type I hyperlipoproteinemia characterized by the accumulation of chylomicrons in the circulation [8]. The validity of our findings is emphazied by previous associations of the S447X variant with a number of lipid parameters [6, 9, 10], and also the presence of the MetS [10], and associations with peak LDL particle size [11]. However, this is the first study, to our knowledge, to associate variants with a pattern of particle sizes created using information from all three fractions of lipoprotein, which associates with the highest IR within the MetS.

The associations observed in this study must be viewed with a number of considerations. These include the small sample size combined with the low MAF for both variants, and the restriction of results to white Americans of European descent. It was this modest sample size that precluded gender stratified analyses. Nonetheless we provide evidence that the LPL 447X allele associates with a pattern of particle sizes (small LDL and HDL, and large VLDL diameters) which, in turn, is associated with the higher IR, WC, fasting TG and glucose and diabetes prevalence among groups where individuals meet ATP-III criteria for the MetS. Given the small sample size, and novelty of the phenotype (lipoprotein diameter pattern) replication in additional studies is warranted.