Original Paper Ann Nutr Metab 2015;66:1–9 DOI: 10.1159/000368425

Received: August 6, 2014 Accepted after revision: September 15, 2014 Published online: December 2, 2014

SH2B1 CpG-SNP Is Associated with Body Weight Reduction in Obese Subjects Following a Dietary Restriction Program Maria Luisa Mansego Fermin Ignacio Milagro Maria Angeles Zulet José Alfredo Martinez CIBERobn, Physiopathology of Obesity and Nutrition, Institute of Health Carlos III, Madrid, and Department of Nutrition, Food Science and Physiology, Nutrition Research Center, University of Navarra, Pamplona, Spain

Abstract The objective of this study was to examine whether 7 SNPs previously associated with obesity-related traits that add or remove potential sites of DNA methylation are accompanied by differential DNA methylation and subsequently affect adiposity variables or body weight reduction in WBC from obese subjects under an energy-restricted program. Material and Methods: Anthropometric measurements were assessed in 47 volunteers recruited within the RESMENA study (Spain). At baseline, DNA from white blood cells was isolated and 7 obesity-related trait CpG-SNPs were genotyped by TaqMan-PCR. Then, methylation levels of CpG-SNP sites were quantified by MassArray® EpiTyperTM or MS-HRM approaches. Results: Differential DNA methylation levels were observed by genotypes in all of the CpG-SNPs analyzed. The FTO and BDNF methylation levels were further correlated with baseline body weight and, BDNF mRNA levels and body weight change, respectively. Moreover, the rs7359397 (SH2B1) was associated with the body weight, body mass index, and truncal fat mass reduction. Conclusions: Our results reveal the interaction of epigenetic and genetic variations in CpG-SNPs, especially in BDNF and SH2B1 genes, and how al-

© 2014 S. Karger AG, Basel 0250–6807/14/0661–0001$39.50/0 E-Mail [email protected] www.karger.com/anm

lele-specific methylation may contribute to elucidate the possible molecular mechanisms as these SNPs are affecting the decrease of mRNA levels and contributing to a lower © 2014 S. Karger AG, Basel body weight reduction.

Introduction

Obesity is the result of the interaction between environmental and behavioral factors, such as diet and exercise, and specific genetic components of each individual [1]. Thus, genetic predisposition is an important factor in the development of obesity, so the genotypic variations in certain nucleotides (single nucleotide polymorphisms, SNPs) associated with obesity may be certainly of great importance explaining this susceptibility [2]. Genomewide association studies (GWAS) have identified numerous loci that are associated with body mass index (BMI), obesity, and even to weight loss after dietary treatment [3, 4]. However, evidence of how these SNPs predispose to obesity remains weak and are available only in certain variants, which mainly affect energy homeostasis and regulation of body weight and appetite [5]. Understanding the molecular mechanisms by which these SNPs have an effect on their target genes is not an easy task, especially because most of them are non-coding SNPs [3]. For J.A. Martínez C/Irunlarrea 1 University of Navarra ES–31008 Navarra, Pamplona (Spain) E-Mail jalfmtz @ unav.es

Downloaded by: St. Andrews University 138.251.14.35 - 12/8/2014 2:21:42 AM

Key Words CpG-SNP · Dietary response · DNA methylation · Obesity · Energy-restricted program

Material and Methods Subjects and Study Protocol The current analysis was conducted within a subsample of 47 subjects from the RESMENA (Metabolic Syndrome Reduction in Navarra) project, a randomized controlled trial [16]. Participants underwent an energy-restricted dietary pattern with a 40/30/30 distribution of macronutrients (carbohydrates/fats/protein) for 8  weeks, as described elsewhere [16]. The study was approved by the Ethics Committee of the University of Navarra (065/2009) and appropriately registered at www.clinicaltrials.gov; NCT01087086. Consequently, all the participants gave written informed consent for  participation in agreement with the Declaration of Helsinki. This work was performed following the CONSORT 2010 guidelines.

2

Ann Nutr Metab 2015;66:1–9 DOI: 10.1159/000368425

Anthropometry and Blood Pressure Anthropometrical measurements (body weight, height, and waist circumference) were conducted according to previously described procedures [16]. Body weight was assessed by using bioimpedance (TANITA SC-330, Tanita, Corporation, Tokyo, Japan) and waist and hip circumferences were measured with a commercial tap following validated protocols, as previously described [16]. Total fat mass and truncal fat mass were evaluated by DXA (Lunar iDXATM, software version 6.0, Madison, Wisc., USA). Systolic and diastolic blood pressures were assessed using a digital monitor (Medisana, MTC, Düsseldorf, Germany) following standardized World Health Organization criteria [17]. Biological Sample and Biochemical Markers Before the weight loss intervention trial, venous blood samples were drawn by venipuncture after a 12-h overnight fast. The EDTA-plasma samples as well as WBC were separated from whole blood by centrifugation at 3,500 r.p.m., 5 ° C, 15 min (Model 5804R, Eppendorf, Germany), and were frozen immediately at –80 ° C until assay (WBC in buffy-coat). Plasma concentrations of glucose and lipid profile were conducted according to previously described procedures [18].  

 

 

 

DNA Isolation Genomic DNA of WBC samples was isolated using the MasterPureTM DNA Purification Kit (Epicentre Biotechnologies, Madison, Wisc., USA) according to the manufacturer’s instructions. DNA was quantified using the PicoGreen® dsDNA Quantitation reagent (Invitrogen, Carlsbad, Calif., USA). Publication Search, SNP Selection, and Identification of CpG-SNPs Initially, SNP selection was based on a search in the PubMed or ISI Web of Knowledge databases of previous reports about association with obesity-related traits in GWAS studies between the years 2009 and 2013. Two search themes were combined using the Boolean operator ‘AND’. The first theme was (‘obesity’ [TIAB] OR ‘BMI’ [TIAB] OR ‘body weight’ OR ‘waist circumference’ OR ‘adiposity’). The second theme was (‘GWA’ OR ‘Genome-Wide Association Study’ [TIAB]). The selection of CpG-SNPs associated with weight loss was only based on candidate gene studies. Exclusion criteria were morbid or extreme obesity, adolescents, or childhood, H-index 1 × 10–5 from the best association test or study without replicates. A total of 18  publications were selected after filtering the records by title, abstract, and inappropriate content. To identify if the selected SNP was CpG-SNP, the DNA sequence of the variant was retrieved from the dbSNP database [19] and was analyzed in order to assess if any of the SNP alleles added or removed a CpG site. From a total of 105 SNPs that were associated with obesity-related trait in GWAS and weight loss in candidate gene studies, 38 (36%) were CpG-SNPs, in which minor alleles that add or remove a CpG site are considered CpG-SNP. After filtering by minor allele frequency (MAF) >0.2, linkage disequilibrium (LD) with another SNP and methylation design problem, a total of 7 CpG-SNPs were selected (table 1). Genotyping Analysis The polymorphisms were genotyped using TaqMan SNP allelic discrimination by means of a 7900HT Fast Real-Time PCR System (Life Technologies, Foster City, Calif., USA). Assay re-

Mansego/Milagro/Zulet/Martinez

Downloaded by: St. Andrews University 138.251.14.35 - 12/8/2014 2:21:42 AM

a better analysis of the pathogenesis of the disease, additional studies that link these SNPs with their biological and molecular functions are required. Moreover, the epigenetic processes, DNA methylation, and histone modification, could influence obesity development, by regulating gene expression [6]. Thus, DNA methylation occurs mainly at a cytosine of a cytosine-phosphate-guanine dinucleotide (CpG) site [7] and increased DNA methylation is generally associated with gene silencing [8]. This modification can be affected by various environmental phenomena, including numerous dietary factors [9]. In this regard, a preliminary study has linked obesity in rats fed with a high-fat diet to increased promoter methylation of leptin [10]. Moreover, several recent studies point to the importance of epigenetic differences between individuals with more or less susceptibility to obesity development or more or less facility for weight loss or maintenance [11, 12]. Furthermore, previous investigations have shown that genetic, epigenetic, and non-genetic factors are integrated to influence the expression of a candidate gene for a particular disease [13, 14]. In this instance, an SNP identified near NDUFB6 gene that introduces a CpG site (CpG-SNP) has been associated with an increased DNA methylation, reduction of gene expression, and insulin resistance in skeletal muscle of elderly patients with diabetes [13]. However, to our knowledge, the role of CpGSNPs associated with obesity in humans has not been examined. Therefore, the objective of this study was to examine whether SNPs previously associated with obesity or weight loss response [3, 4, 15] that add or remove potential sites of DNA methylation are accompanied by differential DNA methylation and subsequently affect adiposity variables or weight loss response in WBC from obese subjects under an energy-restricted program.

Table 1. CpG-SNPs associated with obesity-related traits previously identified by GWAS SNP ID

Candidate Risk MA and SNP gene allele MAF1 position2

Trait

CpG-SNPs where minor allele add a CpG rs1800592 UCP1 C C:0.42 4:141493961 BMI, WLR

Best p value

Reference Taqman Probe for genotyping

1.0×10–3 Nagai et al.

CpG-SNPs where minor allele remove a CpG rs713586 POMC/ C T:0.45 2:25158008 BMI ADCY3 rs6265 BDNF C T:0.23 11:27679916 BMI rs7359397 SH2B1

T

C:0.22

rs7553007 CRP

G

A:0.32

rs9937053 FTO

A

A:0.35

rs4537545 IL6R

T

T:0.39

5.0×10–20 Speliotes et al. 3.6×10–13 Wen et al. 16:28885659 BMI 9.5×10–20 Speliotes et al. 1:159698549 CHD 8.0×10–44 Elliott et al. 16:53799507 Obesity 1.6×10–11 Wang et al. 1:154418879 CHD 2.0×10–14 Elliott et al.

Primers for methylation analysis

C___8866368_20 F-GAGTGATTTTTAAGGGTTATAGGAGAA R-AATATAAAACACTATTCAAATACAAAACAC C___7862140_10 F-TGTTGTTATAGTTTGTGAGATTTTATTGG R-TCCCACTAATATCATTAATCCCTCC C__11592758_10 F-TGGTTGTATGAAGGTTGTTTTTATG R-AAAAAAAACTCCAAAAACACTTAAC C__29242657_10 F-TGATGGTTATGGTTATTGTTTGTTG R-ATCCCTCCAACCTAACTTTAAAAAA C__26627342_10 F-TTTGGAAGTTTAAGGTGGTAA R-ATATCACATATAAAACCTTAACAACA C__29910458_10 F-GGTGGTGGTAGGTGTTTGTAATTT R-AACCACTAACCACCTCTATTCCTTC C__11258666_20 F-TTTTTGTATAATTGGAAGAGTGGATG R-CTCTAAATTTCAAAACATTTTCTTATC

1 Data were obtained from 1000 Genomes. 2 Build 37.1. MA = Minor allele; MAF = minor allele frequency; BMI = body mass index; WLR = weight loss response; CHD = coronary heart disease.

DNA Methylation Levels Analysis The methylation levels of the selected CpG-SNP were performed by a MALDI-TOF mass spectrometry-based method, Sequenom MassArray® EpiTyperTM approach (Sequenom, Calif., USA) after designing specific primers. Bisulfite-treated genomic DNA was amplified using two pairs of primers (table 1). The mean percentage of methylation across each amplicon was calculated for each duplicate of CpG-SNP for each individual. In the case of rs7553007 CpG-SNP for CRP gene, methylationsensitive, high-resolution melting (MS-HRM) was used to quantify the methylation levels due to the failure of the assay designed by MassARRAY system. MS-HRM is based on the different melting profiles of unmethylated and methylated PCR products, due to their different sequence composition (CG content) [20]. The rs7553007 region of the consensus CRP sequence (GenBank #: NT_004487) was used to design primer sets with Primer3 website [21] (table 1). Gene Expression Analysis Twenty four of 47 RNAs were purified from WBC using a TRIzol RNA Isolation Protocol (Invitrogen, Carlsbad, Calif., USA) and the integrity of isolated RNA was also evaluated by Experion (Biorad, Hercules, Calif., USA), following the manufacturer’s instructions. A total of 500 ng of starting material were used as input for the Illumina TotalPrep Amplification Kit Protocol (Life Technologies, Foster City, Calif., USA). Array-based specific gene expression analysis was performed with the HumanHT-12 v4 Expression BeadChip Kit (Illumina)

CpG-SNPs and Body Weight Reduction

and scanned using the Illumina HiScanTM SQ platform (Illumina, San Diego, Calif., USA). The Illumina GenomeStudio Gene Expression Software Module (v 1.9.0) was used to extract the signal intensities of each gene expression probe located within a distance of 500 kb downstream and upstream from the CpG-SNP, which were further preprocessed using the lumi package of Bioconductor in R [22]. Expression microarray data are available in the ArrayExpress database [23] under accession number E-MTAB-2604. Statistical Methods Data are expressed as means (standard deviations, SD), except as otherwise indicated. The Hardy-Weinberg equilibrium (HWE) was tested by the χ2 test for each polymorphism. The volunteers were considered ‘high responders’ (≥8% weight loss from baseline weight after 8 week of energy-restricted intervention; n = 29) and ‘low responders’ (those not achieving a successful weight loss; n = 18), respectively. Baseline and changes of characteristics of participants according to high and low responders were compared. The association between SNPs and CpG methylation, mRNA expression, or anthropometrical traits were based on an additive genetic model, except as otherwise indicated, and p values were adjusted for multiple testing using Bonferroni correction and significance threshold