Baseline anthropometric and metabolic characteristics of obese subjects

Blood samples were collected from the children after an overnight fast to determine plasma glucose, insulin, lipid profile (total cholesterol, triglycerides, low-density lipoprotein, high-density lipoprotein, and apolipoprotein B and A-1), ferritin and leptin by routine laboratory procedures as described previously . Insulin sensitivity was evaluated using homeostasis model assessment (HOMA), a surrogate measure of insulin sensitivity validated in children . Anthropometric data (Tanner stage, weight, height) were collected. Body fat mass was measured by dual-energy X-ray absorptiometry with a total-body scanner (QDR 2000-HOLOGIC, Waltham, MA, USA). Percentage body fat was calculated as 100 × total body fat mass/total body mass. To evaluate the effects of fat distribution, the body was divided into areas corresponding to arms, legs, trunk and head. The trunk region was delineated by an upper horizontal border below the chin, vertical borders lateral to the ribs and a lower border formed by the oblique lines passing through the hip joints. The leg region was defined as the tissue below the oblique lines passing through the hip joints. Body fat distribution was calculated as the ratio of the amount of fat tissue in the trunk region over the amount of fat tissue in the leg region, and was defined as android/gynoid fat mass ratio.

Arterial measurements

A single investigator (YA), who used a real-time B-mode ultrasound imager (Acuson XP 128, Mountain View, CA, USA) performed non-invasive arterial measurements. IMT and lumen diameters during dD and sD were measured at the common carotid artery as described previously . LCSA was calculated as πdD ² /4 and wall cross-sectional area (WCSA) as π(dD/2 + IMT) ² − π(dD/2) ² . Cross-sectional compliance (CSC) and cross-sectional distensibility (CSD) of the common carotid artery were determined from diameter changes during systole and from simultaneously measured pulse pressure (Δ P ), according to the following formulae: CSC = (π[sD ² − dD ² ])/4Δ P (mm ² mmHg ⁻¹ ); and CSD = (sD ² − dD ² )/(dD ² × ΔP) (mmHg ⁻¹ × 10 ⁻² ). Δ P was measured using applanation tonometry during arterial measurements. DWS was calculated as mean arterial pressure × dD/2 × IMT. Incremental elastic modulus ( E _inc ) was calculated as 3 × [1 + LCSA/WCSA])/CSD (mmHg × 10 ² ). In contrast to compliance, which provides information about elasticity of the artery as a hollow structure, the E _inc provides information about the properties of the wall material, independent of the geometry.

After at least 30 min of rest in a recumbent position, arterial endothelial function was studied using a high-resolution vascular ultrasound system. The investigator measured FMD, defined as arterial diameter changes in response to reactive hyperaemia (endothelium-dependent vasodilation induced by a flow increase); and GTNMD, defined as arterial diameter changes in response to the endothelium-independent vasodilator glyceryltrinitrate. Reactive hyperaemia was induced by inflating a blood pressure cuff to 300 mmHg for 4 min; the artery was scanned for 30 s before and 90 s after cuff deflation. A resting scan was recorded 10 m later. Glyceryltrinitrate (400 μg spray) was then given sublingually and the artery was scanned 3 min after the dose.

Selection of polymorphisms and genotyping

With the use of the public databases PubMed ^® ( www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=Pubmed ) and Online Mendelian Inheritance in Man ( www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=OMIM ), we selected variants reported previously to be associated with vascular endothelial dysfunction and artery mechanical property alterations in adults. These polymorphisms are located in genes coding for proteins involved in the renin–angiotensin system (ACE I/D, AGT G-6A and Met235Thr, AGTR1 A1166C), vascular endothelial cell remodelling (MMP3 5A/6A, PAI-1 4G/5G), cell junction and communication (CX37 Pro319Ser), leucocyte adhesion during inflammation (CX3CR1 Val249Ile), arterial vasodilation (NOS3 Glu298Asp), adiponectin secretion (APM1 C-11377G and C-11391G) and lipoprotein metabolism (PON1 Glu192Arg).

All polymorphisms were SNP except for the ACE I/D, which corresponds to the insertion of an Alu-like repetitive sequence into intron 16 . These polymorphisms are located either in the promoter region or in introns or exons and have been shown to cause changes in the function or levels of expression of the encoded protein; the variant allele was reported to occur at least in 5% of the population in previous studies.

The subjects’ DNA was extracted from ethylenediaminetetraacetic acid whole blood samples using the Puregen kit (Gentra Systems, Minneapolis, MN, USA). The SNPs were genotyped with the LightCycler™ (Roche Diagnostics, Basel, Switzerland) based on probes and fluorescence resonance energy transfer between fluorescein and LC Red 640 (Roche Diagnostics). Primers and hybridization probes were designed by TIB MolBiol Syntheselabor (Berlin, Germany). Detailed experimental conditions are available from the authors.

The ACE I/D polymorphism was determined according the technique developed by Marre et al. .

Data analyses

The number of subjects included ( n = 232) was sufficient to detect, with a power of 80%, a 4% variation (threshold based on previous results of our own ) of the flow-mediated dilation trait, defined as the more relevant marker of endothelial function, assuming a recessive genetic model for each polymorphism.

The Chi-square test was applied to check compliance with the Hardy-Weinberg equilibrium of each allele frequency and that no difference in genotype distribution of each polymorphism was present between sexes. Quantitative trait data are expressed as mean ± standard error of the mean. The Gaussian distribution of each variable was tested with the Shapiro-Wilk test. For distributions skewed positively, a log-transformation was applied before the statistical tests but untransformed values are given in tables.

Associations between genotypes (each allele or haplotype alone and combinations of alleles when possible) and phenotype traits were tested with the general linear model analysis of covariance (least-square mean regression) with adjustments on covariates as follows: age, sex, Tanner stage, homeostasis model assessment, ferritin, android/gynoid fat-mass ratio, leptin and lipid profile. Each genotype was assessed with the use of dominant, recessive and additive genetic models.

All statistical analyses were performed with JMP software (SAS Institute Inc, Cary, NC, USA). Statistical significance was defined from p < 0.05.