Facultad de Farmacia y Nutrición

Nutrigenetic approaches for personalized management of obesity and metabolic syndrome traits

Enfoques nutrigenéticos para el tratamiento personalizado de la obesidad y componentes del síndrome metabólico

Leticia Goñi Mateos Pamplona, 2017

Facultad de Farmacia y Nutrición

Memoria presentada por Dña. Leticia Goñi Mateos para aspirar al grado de Doctor por la Universidad de Navarra.

Leticia Goñi Mateos

El presente trabajo ha sido realizado bajo nuestra dirección en el Departamento de Ciencias de la Alimentación y Fisiología de la Facultad de Farmacia y Nutrición de la Universidad de Navarra y autorizamos su presentación ante el Tribunal que lo ha de juzgar.

Pamplona, 14 de septiembre de 2017

VºBº Director VºBº Co-Director Prof. J. Alfredo Martínez Hernández Dr. Marta Cuervo Zapatel

Este trabajo ha sido posible gracias a la financiación de diversas entidades: Gobierno de Navarra (Aditech PT025), Ministerio de Economía y Competitividad (AGL2013-45554-R), Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición (CIBERObn) Instituto de Salud Carlos III (ISCIII), Centro de Investigación en Nutrición (Universidad de Navarra). La investigación que ha dado lugar a estos resultados ha sido impulsada por la beca predoctoral 2012-2014 de la Asociación de Amigos de la Universidad de Navarra y las becas predoctoral 2015-2017 y de movilidad del Ministerio de Educación, Cultura y Deporte.

A mi familia

- AGRADECIMIENTOS / ACKNOWLEDGMENTS -

A continuación, quisiera plasmar mi agradecimiento a todas aquellas personas e instituciones que han hecho que hoy esté escribiendo estas líneas.

A nivel institucional quisiera agradecer a la Universidad de Navarra y a la Facultad de Farmacia y Nutrición que hace ya 10 años decidió acogerme en sus pasillos, aulas y laboratorios. En especial al Departamento de Ciencias de la Alimentación y Fisiología y al Centro de Investigación en Nutrición por darme la oportunidad de realizar este trabajo.

En especial quisiera agradecer al Prof. Alfredo Martínez y a la Dra. Marta Cuervo, los directores de la presente tesis. Mi más sincero agradecimiento a los dos por vuestra atenta supervisión, dedicación, preocupación y confianza puesta en mí. Gracias a los dos por ayudarme en mi formación profesional e investigadora.

Además, quisiera agradecer a la Asociación de Amigos de la Universidad de Navarra (beca predoctoral) y al Ministerio de Educación, Cultura y Deporte (beca predoctoral y de movilidad) por la financiación recibida durante estos años de tesis doctoral. Así como a las instituciones que han hecho posible todos los proyectos recogidos en la presente tesis: CINFA, Gobierno de Navarra (ADITECH), Ministerio de Economía y Competitividad, Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición (CIBERObn) Instituto de Salud Carlos III (ISCIII).

No quisiera dejar de agradecer a todo el personal del Departamento de Ciencias de la Alimentación y Fisiología y/o Centro de Investigación en Nutrición que me han guiado por este camino. En especial y en primer lugar a Fermín por adentrarme en el campo de la genética y la orientación en cada uno de los capítulos de este trabajo. En segundo lugar, a Jose Ignacio, gracias por todos tus consejos siempre acompañados de risas. Y no me olvido de Carlos, Pedro y Marian, gracias por vuestras conversaciones y orientaciones tanto a nivel personal como profesional.

También quiero agradecer a todo el equipo del estudio de intervención nutricional incluyendo al personal de la Unidad Metabólica (Blanca, Salomé, María, Angels, Santiago), del CIMA (Fernando, Lourdes, Beatriz) y a la técnico Ana por dejarme “cacharrear” con tanta paciencia en el laboratorio. En especial a Iosune y Laura con las que he logrado crear un gran equipo y que habéis hecho que todo sea alcanzable de la manera más sencilla. Aprovecho estas líneas también para agradecer a nuestro bioinformático Kike, cuanta estadística tenemos que aprender de ti y con cuanto entusiasmo.

No quisiera olvidarme de todos los alumnos de máster y grado que posiblemente sin ser conscientes han contribuido en esta tesis: Astrid, Maritza, Miguel, Amanda, Liz, Karleny, Carla, María, Javier, Natalia y Valeria.

Gracias a los voluntarios que han participado en cada uno de los estudios aplicados en la presente tesis, quienes tal vez no comprendan claramente que son el centro de nuestro trabajo; sin ellos y sin su esfuerzo este proyecto carece absolutamente de sentido.

Gracias a los compañeros y amigos de la sala de ordenadores, a los que se fueron, a los que siguen estando y a los que acaban de llegar. Gracias por vuestros ánimos durante estos intensos años, como todos sabéis pasamos nuestros momentos menos buenos, pero siempre pesan más los buenos. Os deseo lo mejor. Mil gracias al “Granada Team” por compartir tantos momentos juntas, y los que nos quedan… Gracias porque incluso desde la distancia siempre he contado con vuestro apoyo.

I whish to express my sincere gratitude to Prof. Lu Qi for having given me the opportunity to work in her team at the Tulane University. Thank you to all her team for the warm welcome. Many thanks for the kind help to Danny and Yoriko.

Gracias a mi amig@s por estar siempre conmigo y preguntarme por mis avances y como no, por cuando termino de estudiar! Muy en especial a mi mejor amiga Itziar, por compartir tantos momentos juntas, escuharme y aconsejarme.

A Joseba, quien ha sido mi mayor apoyo en todos estos años de tesis. Gracias por esos momentos de locura y desorden que me han permitido desconectar de mi vida profesional y que tanto necesito en mi ordenada vida. Gracias por comprometerte a estar a mi lado y hacerme feliz siempre.

Gracias muy especialmente a mi familia. A mis abuelos, tíos y primos quienes sin saber muy bien a que me dedico siempre han estado pendientes de mí. Gracias a mis hermanos Victor y Sara simplemente por estar ahí siempre y también a sus “apegados”. GRACIAS a mis padres, a quienes les dedico este trabajo y les agradezco su apoyo incondicional, escuchándome siempre que lo he necesitado y apoyándome en todas mis decisiones. Por enseñarme a mí y a mis hermanos que el esfuerzo y la constancia permiten alcanzar todo aquello que nos proponemos.

GRACIAS - THANK YOU - ESKERRIK ASKO - OBRIGADO --

- LIST OF ABBREVIATIONS -

ADA American Association AGEN Asian Genetic Epidemiology Network ANCOVA Analysis of the covariance ANOVA Analysis of the variance AUC Area under the curve BIA Bioimpedance BMI Body mass index CARe Candidate Association Resource CH Carbohydrates CT Computed tomography DBP Diastolic blood pressure DXA Dual-energy-X-ray absorptiometry DZ Dizygotic EPIC European Prospective Investigation into Cancer ESC European Society of Cardiology ESH European Society of Hypertension FFQ Food frequency questionnaire FGFQ Food groups frequency questionnaire GIANT Genetic Investigation of Anthropometric Traits Global BPgen Global Blood Pressure Genetics GRS Genetic risk-allele score GWAS Genome wide association study Hb1Ac Glycated hemoglobin HDL High density lipoprotein HPFS Health Professional’s Follow-up Study HWE Hardy-Weinberg equilibrium ICBP International Consortium of Blood Pressure ICC Intraclass correlation coefficient IDF International Diabetes Federation IDL Intermediate density lipoprotein IFG Impaired fasting glucose IGT Impaired glucose tolerance IR resistance JDRF Juvenile Diabetes Research Foundation JNC7 7th report of the Joint National Committee LARS Least angle regression LDL Low density lipoprotein LOA Limits of agreement MAGIC Meta-Analysis of Glucose and Insulin-Related Traits Consortium

MODY Maturity onset diabetes of the young MUFA Monounsaturated fatty acids MZ Monozygotic NDM Neonatal diabetes mellitus NHS The Nurses’ Health Study OGTT Oral glucose tolerance test PA Physical activity POUNDS Lost Preventing Overweight Using Novel Dietary Strategies PUFA Polyunsaturated fatty acids SAT Subcutaneous adipose tissue SBP Systolic blood pressure SD Standard deviation SEEDO Spanish Society for the Study of Obesity SFA Saturated fatty acids SNP Single nucleotide polymorphism VAT Visceral adipose tissue VLDL Very low density lipoprotein WAGR Wilm’s tumor aniridia genitourinary anomalies and mental retardation WGHS Women Genome Health Study WHO World Health Organization

- LIST OF ABBREVIATIONS OF –

AANAT Aralkylamine N-acetyltransferase ABCC8 ATP binding cassette subfamily C member 8 ACE Angiotensin-converting enzyme ACSL5 Acyl-CoA synthase long-chain family member 5 ADAM17 ADAM metallopeptidase domain 17 ADAMTS9 ADAM metallopeptidase with thrombospondin type 1 motif 9 ADAMTSL3 ADAMTS like 3 ADCY Adenylate cyclase ADIPOR1 Adiponectin receptor 1 ADIPOQ Adiponectin C1Q and collagen domain containing ADRB Adrenoreceptor beta AGBL4 ATP/GTP binding protein like 4 AGT Angiotensinogen AGTR1 Angiotensin II receptor type 1 ALDH2 Aldehyde dehydrogenase 2 family APO Apolipoprotein ARAP1 ArfGAP with RhoGAP domain Ankyrin repeat and PH domain BDNF Brain derived neurotrophic factor BMP2 Bone morphogenetic protein 2 C6orf106 6 open reading frame 106 CADM Cell adhesion molecule CAV2 Caveolin 2 CCK Cholecystokinin CCDC92 Coiled-coil domain containing 92 CD300LG CD300 molecule like family member g CDKAL1 CDK5 regulatory subunit associated protein 1 like 1 CDKN2A/B Cyclin dependent kinase inhibitor 2A/2B CELSR2 Cadherin EGF LAG seven-pass G-type receptor 2 CETP Cholesteryl ester transfer protein CLOCK Circadian regulator CNPY2 Canopy FGF signaling regulator 2 CNR2 Cannabinoid receptor 2 COBLL1 Cordon-bleau WH2 repeat protein like 2 COLEC-10-MAL2 Collectin subfamily member 10-mal T-cell differentiation protein 2 CRTC1 CREB regulate transcription coactivator 1 CTNNBL1 Catenin beta like 1 CTSS Cathepsin S CYB5B Cytochrome b5 type B CYP11B2 Cytochrome P450 family 11 subfamily B member 2 DNAJC17 DNaJ heat shock protein family (Hsp40) member C17

DRD2 Dopamine receptor D2 EFEMP1 EGF containing fibulin like like extracellular matrix protein 1 ERB34 Erb-b2 receptor tyrosine kinase 4 ETV5 ETS variant 5 FAAH Fatty acid amide hydrolase FABP Fatty acid binding protein FANCL Fanconi anemia complementation group L FGF21 Fibroblast growth factor FHIT Fragile histidine triad FLJ33534 Uncharacterized LOC285150 FTO Fat mass and obesity associated GALNT10 Polypeptide N-acetylgalactosaminyltransferase 10 GCK Glucokinase GCKR Glucokinase regulatory protein GHRL Ghrelin and obestatin prepropeptide GIPR Gastric inhibitory polypeptide receptor GNAS GNAS complex GNAT2 G protein subunit alpha transducin 2 GNPDA2 Glucosamine-6-phosphatase deaminase 2 GPRC5B G protein-coupled receptor class C group 5 member B GRAMD3 GRAM domain containing 3 GRB14 Growth factor receptor bound protein GRID1 Glutamate ionotropic receptor delta type subunit 1 GSDMB Gardermin B HHEX Hematopoietically expressed HIF1AN Hypoxia inducible factor 1 alpha subunit inhibitor HIP1 Huntingtin interacting protein 1 HLA-DRB5 Major histocompatibility complex, class II, DR beta 5 HMGCR 3-hydroxy-3-methylglutaryl CoA HNF4 Hepatocyte nuclear factor 4 HS6ST3 Heparin sulfate 6-O-sulfatransferase 3 HSD17B4 Hydroxysteroid 17-beta dehydrogenase 4 HTR2C 5-hydroxytryptamine (serotonin) receptor 2C IGF2BP Insulin like growth factor 2 mRNA binding protein IGFBP3 Insulin like growth factor binding protein 3 INS Insulin INSIG2 Insulin induced gene 2 INSR Insulin receptor IRS1 Insulin receptor substrate 1 ITIH4 Inter-alpha-trypsin inhibitor chain family member 4 KCNJ11 Potassium voltage-gated channel subfamily J member 11

KCTD15 Potassium channel tetramerization domain containing 15 KLHL Kelch like family member KCNQ1 Voltage-gated channel subfamily Q member 1 KLF14 Kruppel like factor 14 KLHL32 Kelch like family member 32 LDLR Low-density lipoprotein receptor LEKR1 Leucine, glutamate and lysine rich 1 LEP Leptin LEPR Leptin receptor LHX2 LIM homeobox 2 LIPC Lipase c hepatic type LPIN Lipin 1 LPL Lipoprotein lipase LRP LDL receptor related protein LRRN6C Leucine rich repeat neuronal 6C LY86 Lymphocyte antigen LYPLAL1 Lysophospholipase like 1 MAP2K5 Mitogen-activated protein kinase kinase 5 MAP3K1 Miogen-activated protein kinase kinase kinase 1 MC4R Melanocortin 4 receptor MECOM MDS1 and EVI1 complex locus MIR148A MicroRNA 148a MLXIPL MLX interacting protein like MN1 MN1 proto-oncogene transcriptional regulator MRPS33P4 Mitochondrial ribosomal protein S33 pseudogene 4 MSRA Methionine sulfoxide reductase A MTCH2 Mitochondrial carrier homolog 2 MTIF3 Mitochondrial translational initiation factor 3 MTNR1B Melatonin receptor 1B MYL2 Myosin light chain 2 NAV1 Neuron navigator 1 NCP1 Niemann-Pick Disease, type C1 NEGR1 Neuronal growth regulator 1 NFE2L3 Nuclear factor erythroid 2 like 3 NOS3 Nitric oxide synthase NID2 Nidogen 2 NPC1 NPC intracellular cholesterol transporter 1 NRXN3 Neurexin 3 NT5C2 5’-nucleotidase, cytosolic II NTRK2 Neurotrophic tyrosine kinase receptor type 2 NUDT3 Nudix hydrolase 3

PCSK Proprotein convertase subtilisin/kexine type PEPD Peptidase D PIK3R1 Phosphoinositide-3-kinase regulatory PLA2G6 Phospholipase A2 group VI PLAUR Plasminogen activator, urokinase receptor PLIN Perilipin POMC Pro-opiomelanocortin PPAR Peroxisome proliferator activated receptor PPARGC1A PPARG coactivator 1 alpha PPM1K Protein phosphatase Mg2+/Mn2+ dependent 1K PRKD1 Protein kinase D1 PTBP2 Polypyrimidine tract binding protein 2 PTN Pleiotrophin QPCTL Glutaminyl-peptide cycotransferase like RASA2 RAS p21 protein activator 2 REN Renin RPL27A Ribosomal protein L27a RPTOR Regulatory associated protein of MTOR complex 1 RREB1 Ras responsive element binding protein RSPO3 R-spondin 3 SEC16B SEC16 homolog B, endoplasmic reticulum export factor SH2B1 SH2B adaptor protein 1 SHAS1 SAM and SH3 domain containing 1 SIM1 Single-minded homolog 1 SLC_A_ Solute carrier family _ member _ SOS1 SOS Rac/Ras guanine nucleotide exchange factor 1 SPRY2 Sprouty RTK signaling antagonist 2 STXBP6 Syntaxin binding protein 6 TBX15 T-box 15 TCF7L2 7 like 2 TFAP2B Transcription factor AP-2 beta TM6SF2 Transmembrane 6 superfamily member 2 TMEM Transmembrane protein TNNI3K TNNI3 interacting kinase TOMM40 Trnaslocase of outer mitochondrial membrane 40 TOX2 TOX high mobility group box family member 2 TRAPPC9 Trafficking protein particle complex 9 TUFM Tu translocation elongation factor, mitochondrial UBE2E2 Ubiquitin conjugating enzyme E2E2 UCP Uncoupling protein USF1 Upstream regulatory factor 1

VEGFA Vascular endothelial growth factor WARS2 Tryptophanil tRNA synthetase 2 mitochondrial WWOX Ww domain containing oxidoreductase ZNF608 Zinc finger protein 608 ZZZ3 Zin finger type containing 3

- TABLE OF CONTENTS -

INTRODUCTION ...... 1 1. OBESITY: DEFINITION, CAUSES AND PREVALENCE ...... 3 2. GENETICS OF OBESITY ...... 4 2.1. Evidence for a genetic component of obesity ...... 5 2.2. Monogenic obesity ...... 6 2.3. Polygenic obesity ...... 7 3. GENETICS OF BODY WEIGHT REGULATION ...... 21 3.1. Genes involved in energy expenditure ...... 22 3.2. Genes involved in appetite control and food intake ...... 22 3.3. Genes involved in adipogenesis and lipid metabolism ...... 23 3.4. Other polymorphisms of interest ...... 24 3.5. Gene-diet interactions on body weight regulation ...... 24 4. GENETICS OF OBESITY METABOLIC RELATED TRAITS ...... 26 4.1. Hyperlipidemia ...... 26 4.2. ...... 31 4.3. Hypertension ...... 37

HYPOTHESIS AND OBJECTIVES ...... 43 1. HYPOTHESIS ...... 45 2. GENERAL OBJECTIVE ...... 45 3. SPECIFIC OBJECTIVES ...... 45

SUBJECTS AND METHODS ...... 47 1. VALIDATION STUDY OF A FOOD GROUPS FREQUENCY QUESTIONNAIRE ...... 50 1.1. Study population ...... 50 1.2. Data collection ...... 51 2. NUTRIGENETIC SERVICE NS COHORT ...... 52 2.1. Study population ...... 53 2.2. Intervention ...... 53 2.3. Data collection ...... 53 2.4. Genotyping ...... 54 3. OBEKIT STUDY ...... 55 3.1. Study population ...... 55 3.2. Intervention ...... 57 3.3. Data collection ...... 57

3.4. Genotyping ...... 59 4. POUNDS LOST TRIAL ...... 59 4.1. Study population ...... 59 4.2. Intervention ...... 61 4.3. Data collection ...... 61 4.4. Genotyping ...... 63 5. NUGENOB STUDY ...... 63 5.1. Study population ...... 63 5.2. Intervention ...... 64 5.3. Data collection ...... 65 5.4. Genotyping ...... 66 6. STATISTICAL ANALYSES ...... 66

RESULTS ...... 69

INTRODUCTORY RESEARCH ...... 71 CHAPTER 1. Single-nucleotide polymorphisms and DNA methylation markers associated with central obesity and regulation of body weight ...... 73 CHAPTER 2. Future perspectives of personalized weight loss interventions based on nutrigenetic, epigenetic, and metagenomic data ...... 113 CHAPTER 3. Validation of a food groups frequency questionnaire based in an exchange system ...... 135

ORIGINAL RESEARCH ...... 159 CHAPTER 4. A genetic risk tool for obesity predisposition assessment and personalized nutrition implementation based on macronutrient intake ...... 161 CHAPTER 5. Significant phenotype and genotype predictors of BMI in adult population ... 191 CHAPTER 6. Interaction between an ADCY3 genetic variant and two weight lowering diets affecting body fatness and body composition outcomes depending on macronutrient composition: a randomized trial ...... 213 CHAPTER 7. Gene-gene interplay and gene-diet interactions involving the MTNR1B rs10830963 variant with body weight loss ...... 231 CHAPTER 8. Circadian rhythms related MTNR1B genetic variant modulates the effect of weight-loss diets on changes in adiposity and body composition: The POUNDS Lost trial . 253 CHAPTER 9. Macronutrien-specific effect of the MTNR1B genotype on lipid levels in response to 2-year weight-loss diets ...... 271

CHAPTER 10. Effect of the interactions between diet composition and the PPM1K genetic variant on and b cell function markers during weight loss: results from the NUGENOB randomized trial ...... 289 CHAPTER 11. Influence of fat intake and BMI on the association of rs17999983 NOS3 polymorphism with blood pressure levels in an Iberian population ...... 311

GENERAL DISCUSSION ...... 331 1. JUSTIFICATION OF THE STUDY ...... 333 2. GENETIC SUSCEPTIBILITY TO OBESITY ...... 333 3. GENETIC VARIANTS IMPLICATED IN WEIGHT LOSS ...... 337 4. GENETIC VARIANTS IMPLICATED IN OBESITY RELATED TRAITS ...... 341 5. STRENGTHS AND LIMITATIONS ...... 346 6. COROLLARY ...... 347

CONCLUSIONS ...... 349

REFERENCES ...... 353

INTRODUCTION

Introduction

1. OBESITY: DEFINITION, CAUSES AND PREVALENCE

Obesity has been defined, by The World Health Organization (WHO), as an abnormal or excessive fat deposition that contributes to increase morbidity and mortality (1). This enlarged fat accumulation is mainly stored as triglycerides in the adipose tissue (subcutaneous and intra-abdominal accumulation), which constitutes a long-term energy reservoir (2).

Generalizing, obesity is attributed to a chronic positive energy balance; arising, when individuals consume more energy than they expend and is maintained over time (3). Several factors could affect the energy equation such as lifestyle behaviors (dietary habits, sedentarism, sleep duration) (4–6), social determinants (education level, economic status) (7,8), endocrine disorders (hypothyroidism, Cushing’s syndrome) (9) or prescription of certain medications (β-blockers, corticosteroids, antipsychotics) (10). Moreover, the fact that not all individuals exposed to the same environmental risk factors develop obesity, lends support to an underlying genetic (11), epigenetic (12) and metagenomic (13) components to the disorder (Figure 1). ­ Familyhistory educationlevel Exercise Stress

Disturbed sleep

Medication

Endocrine disorder

Disturbed sleep

↓socioeconomic level

Figure 1. Physiopathology of obesity. Adapted from: Franks et al. Diabetes Care, 2013 (14)

Obesity diagnosis should be the first step toward treatment (15). Nowadays, in clinical practice, the most used tool to detect overweight and obesity in middle-aged adults is the body mass index (BMI), calculated dividing body weight (kg) by the square of height (m). The WHO (16) and the Spanish Society for the Study of Obesity (SEEDO) (17) have established the following criteria to classify the nutritional status of the individuals (Table 1).

3 Introduction

Table 1. The WHO and SEEDO classifications of the nutritional status by the BMI BMI WHO (16) SEEDO (17) <18.5 Underweight Underweight 18.5-24.9 Normal weight Normal weight 25 – 26.9 Overweight level I Overweight (pre-obesity) 27 – 29.9 Overweight level II (pre-obesity) 30 – 34.9 Obesity class I Obesity class I (moderately obese) 35 – 39.9 Obesity class II Obesity class II (severely obese) 40 – 49.9 Obesity class III (very severely obese) Obesity class III ≥ 50 Obesity class IV (extreme obese) Abbreviations: BMI, Body mass index; SEEDO, Spanish Society for the Study of Obesity; WHO, World Health Organization

However, BMI makes an assumption about the distribution of adipose tissue and lean muscle mass of an individual using only weight and height, over or underestimating adiposity depending on subject complexion, age or ethnicity (18,19). Thus, measures to evaluate fat mass have been applied such as underwater weighing method, air-displacement plethysmography and dual-energy-X-ray absorptiometry (DXA) (20). However, the high cost or complexity of such equipment, limit their use and simple methods such as skinfolds and bioelectrical impedance analysis are using in both, clinical and research field (21). The SEEDO has proposed a percentage of body fat threshold for obesity definition of >25% and >33% for men and women, respectively (22).

The prevalence of overweight and obesity is rising steadily not only in high-income countries but also in low-income countries reaching epidemic proportions (1). In fact, the proportion of obese subjects has almost doubled from 6.4% in 1980 to 12.0% in 2008, with half of this increase occurring between 2000 and 2008 (23).

According to the WHO, more than 1.9 billion adults were overweight and over 600 million of them were obese in 2014 (1). In percentage, 39% of the population was overweight and 13% was obese. Although obesity is a worldwide problem there are differences in overweight and obesity levels and tendencies among countries (24). Current trends suggest that 3.3 billion adults (57.8% of the population) will be overweight or obese by 2030 (25).

2. GENETICS OF OBESITY

Due to the rapid change in the prevalence of obesity several theories have been hypothesized to explain this phenomenon. In 1962, Neel proposed the “thrifty gene hypothesis” and cited with increasing prevalence of obesity (26,27). This theory asserts that in prehistoric times there would have been positive selection for traits conferring the ability to store energy efficiently in periods of limited food availability (26). Thrifty genes are those genes that enable to efficiently collect and process food to deposit fat during periods of abundance in order to

4 Introduction provide for periods of food shortage (feast and famine). However, in modern society, such genes are disadvantageous because of the constant abundance of food, and the result is widespread obesity. As an alternative to the “thrifty gene hypothesis”, Speakman proposed the “drifty gene hypothesis” (28). This theory, in contrast to the “thrifty genotype”, proposes that genes predisposing to obesity were not under any positive selection, but have rather been subject to random drift because of an absence of selection (28,29). In light of these theories clearly the role of genetics in modifying obesity risk in a world of nutrient excess in detrimental of physical activity is complex and far from being completely understood.

2.1. Evidence for a genetic component of obesity

Compelling data from heritability studies, which are based on the fact that monozygotic twins are genetically identical, while dizygotic twins share only 50% of their genetic background, have showed that genetic component plays an important role in an individual’s predisposition of becoming obese (30). A previous review of 32 twin studies reported that the mean BMI heritability was 73%, although the BMI heritability ranged from 31% to 90% (31). The heterogeneity of BMI heritability derived from twin studies could be attributable to different factors. Thus, one study concluded that the heritability of BMI increased from 77% to 84% over 25 years of follow-up, showing that BMI heritability is sensitive to age (32). This finding has been ascertained in a meta-analysis which involved 8,527 monozygotic and 8,317 dizygotic twin pairs (33). Another study demonstrated that not only the time period of observation, but also the socioeconomic status influenced the heritability proportion of BMI (31).

Adoption and familial studies also provide evidence of the genetic contribution to obesity. Stunkard et al. (1986) showed a strong correlation between the adoptees and biological parents with respect to BMI but not with adoptive parents (32). In addition, Sorensen et al. (1989) reported the association regarding BMI between adoptees and their biological full and half siblings who were reared separately by the biological parents of the adoptees (34). Although a BMI association existed between adoptees and their half siblings, the BMI of the adoptees was found to be most closely related to the fatness of the full siblings.

Further evidence comes from differences in prevalence of obesity among racial/ethnic groups. For example, the prevalence of obesity among Pima Indians is 50% or more, meanwhile among Caucasian and Asian populations is around 35% (35).

5 Introduction

2.2. Monogenic obesity

The term monogenic obesity refers to a number of rare forms of severe obesity resulting from mutations of large effect size in a single gene or a chromosomal region, that affect about 5% of the obese population (36). There are two major forms of monogenic obesity: non-syndromic and syndromic, both characterized, ingeneral, by severe and early onset obesity.

In non-syndromic forms, extreme obesity is the predominant feature accompanied by other metabolic disorders (37). There are at least eight well known genes mutations in monogenic non-syndromic form of obesity related to the leptin-melanocortin signaling pathway present in the hypothalamus and therefore, affect regulation of food intake and energy expenditure (Table 2): leptin (LEP), leptin receptor (LEPR), pro-opiomelanocortin (POMC), proconvertase 1 (PCSK1), melanocortin 4 receptor (MC4R), brain-derived neurotrophic factor (BDNF), neurotrophic tyrosine kinase receptor type 2 (NTRK2) and single-minded homolog 1 (SIM1) (38,39).

Table 2. Non-syndromic forms of obesity Gene Chr location Mutations Obesity phenotype

LEP 7q32.1 DG133, Arg105Trp Extreme, early-onset obesity, hyperphagia LEPR 1p31.3 Exon 16 splice donor G àA Extreme, early-onset obesity, hyperphagia POMC 2p23.3 G7013T, 7133delC, C3804A, A6851T, Early-onset obesity 6906delC, 6996del, 7100insGG, 7134delG PCSK1 5q15 Gly483Arg, AàC +4 intron 5 donor Childhood-onset obesity, elevated proinsulin, splice site, Glu250Stop, Del213Ala hypocortisolemia, depressed POMC, reactive MC4R 18q21.32 >150 Early-onset obesity, hyperphagia, increased fat mass, increased lean mass BDNF 11p13 46, XX, inv(11) (913915.3) Severe obesity, hyperphagia, body weight NTRK2 9q22.1 Y722C Severe early-onset obesity, hyperphagia SIM1 6q16.3 de novo balanced translocation Early-onset obesity, hypotonia, developmental delay 1p22.1 and 6q16.2 Abbreviations: LEP, Leptin; LEPR, Leptin receptor; POMC, Pro-opiomelanocortin; PCSK1, Proconvertase 1; MC4R, Melanocortin 4 receptor; BDNF, Brain-derived neurotrophic factor; NTRK2, Neurotrophic tyrosine kinase receptor type 2; SIM1, Single-minded homolog 1; Chr, Chromosome Adapted from: Albuquerque D et al. Mol Genet Genomics, 2015 (37)

Syndromic forms of obesity refer to obesity cases that occur associated with different clinical phenotypes such as mental retardation or organ-specific developmental abnormalities (37). The most common forms of syndromic obesity are Wilm’s tumor aniridia genitourinary anomalies and mental retardation (WAGR), Prader-Willi, Bardet-Bield, and Alström and Cohen syndromes (Table 3) (11,36).

6 Introduction

Table 3. Syndromic forms of obesity Syndrome Gene Chr location Obesity phenotype

WAGR BDNF 11p14.1 Obesity Prader Willi Contiguous gene disorder 15q11-13 Neonatal hypotonia, poor feeding evolving into extreme hyperphagia, central obesity Bardet-Biedl BBS1-BBS12 11q13.2 Progressive late childhood obesity Alström ALMS1 2p13.1 Mild truncal obesity Cohen VPS13B 8q22 Mild truncal obesity

Abbreviations: WAGR, Wilm’s tumor aniridia genitourinary anomalies and mental retardation; BDNF, BDNF, Brain-derived neurotrophic factor; BBS, Bardet-Briedl syndrome; ALMS1, Alström syndrome 1; VPS13B, Vacular protein sorting 13 homolog B; Chr, Chromosome Adapted from: Albuquerque et al. Mol Genet Genomics, 2015 (37)

2.3. Polygenic obesity

Polygenic obesity, also known as “common obesity”, results from the combined effect of several genetic variants in multiple genes (40). In contrast with monogenic obesity, in polygenic obesity each polymorphism has a small effect to contribute to obesity predisposition. Therefore, each genetic variant additionally requires the presence of other variants and an obesogenic environment to develop the obese phenotype (41). There are two main approaches in the identification of single nucleotide polymorphisms (SNPs) associated to obesity: candidate gene analyses and genome wide association studies (GWASs).

2.3.1. Candidate gene studies

First studies analyzing the genetic basis of common obesity relied on candidate gene studies which attempts to detect association between genetic markers in predefined genes considered to be candidates for the phenotype of interest (37). Most candidate gene approaches looked at genetic markers previously implicated in monogenic forms of obesity or reported in animal studies (42).

Several hundred obesity candidate gene studies have been carried out, with 127 identified genes related to obesity or obesity associated phenotypes until the last update in 2005 of the Human Obesity Gene Map (43). The most well-known obesity candidate gene associations include peroxisome proliferator-activated receptor gamma (PPARG), adrenoceptor beta 2 and 3 (ADRB2 and ADBR3), LEP and LEPR, uncoupling protein 1, 2 and 3 (UCP1, UCP2 and UCP3), and MC4R (8).

Although candidate gene studies continue to contribute to our understanding of the genetic basis of obesity, some studies have achieved limited success in identifying susceptibility variants due to the small impact of them in the phenotype. Moreover, conflicting results have

7 Introduction been reported due to limit study power, sample size, phenotyping and inclusion criteria, among other factors. Thus, some authors have proposed to meta-analyze data from different association studies to achieve statistical power to define if the polymorphism influences or not the obesity susceptibility (Table 4).

Table 4. Summary of meta-analysis of candidate gene studies in adult populations n (cases/ Reference Gene Polymorphism Outcome controls)

Bender et al., 2011 (44) LEPR rs1137101 NA Associated with obesity risk rs1137100 NA No associated with obesity risk rs8179183 NA No associated with obesity risk Peng et al., 2011 (45) FTO rs9939609 41,374/ 69,837 Associated with obesity risk rs1421085 41,374/ 69,837 Associated with obesity risk rs8050136 41,374/ 69,837 Associated with obesity risk rs17817449 41,374/ 69,837 Associated with obesity risk rs1121980 41,374/ 69,837 Associated with obesity risk Wang et al., 2011 (46) ENPP-1 rs1044498 11,372/ 12,952 Associated with obesity risk Zhao et al., 2011 (47) FABP2 rs1799883 10,974 No associated with BMI Xi et al., 2012 (48) MC4R rs17782313 48,413/ 134,392 Associated with obesity risk rs12970134 19,238/ 31,913 Associated with obesity risk rs571312 18,060/ 59,573 Associated with obesity risk rs17700144 2,880/ 7,880 Associated with obesity risk rs4450508 5,609/ 11,319 Associated with obesity risk Yu et al., 2012 (49) ADIPOQ rs1501299 1,329/ 1,225 Associated with obesity risk IL6 rs1800795 923/ 666 Associated with obesity risk LEP rs7799039 1,002/ 1,418 No associated with obesity risk LEPR rs1137101 3,026/ 3,312 No associated with obesity risk RETN rs1862513 4,098/ 5,762 No associated with obesity risk TNFα rs1800629 1,918/ 2,302 Associated with obesity risk Galbete et al., 2013 (50) PPARG rs1801282 9,286/ 39,806 Associated with obesity risk Brondani et al., 2014 (51) UCP1 rs1800592 NA No associated with BMI Stijnen et al., 2014 (52) PCSK1 rs6232 89,047/ 127,367 Associated with obesity risk and BMI rs6234-rs6235 89,047/ 127,367 Associated with obesity risk, BMI and WC Zhang et al., 2014 (53) ADRB2 rs1042714 3,285/ 6,591 Associated with obesity risk Zhang et al., 2014 (54) LEP rs7799039 1,372/ 1,616 Overall population no associated with obesity risk. Associated with obesity risk among subjects of mixed race from South America Zhang et al., 2014 (55) UCP2 rs660339 2,011/ 2,752 Associated with overweight-obesity risk rs659366 5,780/ 10,196 Associated with overweight-obesity risk Mansoori et al., 2015 (56) PPARG rs1801282 39,647 Associated with BMI Yan et al., 2015 (57) LEP rs7799039 1,235/ 1,359 Overall population no associated with obesity risk. Associated with obesity risk among African individuals Yao et al., 2015 (58) PPARG rs1801282 6,491/ 8,242 Associated with obesity risk Lin et al., 2016 (59) GNB3 rs5443 3,171/ 7,225 Associated with overweight-obesity risk

Abbreviations: LEPR, Leptin receptor; FTO, Fat mass and obesity associated; ENPP-1, Ectonucleotide pyrophosphatase/phosphodiesterase 1; FABP2, Fatty acid binding protein 2; MC4R, Melanocortin 4 receptor; ADIPOQ, Adiponectin C1Q and collagen domain containing; IL6, Interleukin 6; LEP, Leptin; RETN, Resistin; TNFα, Tumor necrosis factor α; PPARG, Peroxisome proliferator activated receptor gamma; UCP1, Uncoupling protein 1; PCSK1, Proconvertase 1; ADRB2, Adrenoreceptor beta 2; UCP2,Uncoupling protein 2; GNB3, G protein subunit beta 3; NA, No available; BMI, Body mass index; WC, Waist circumference

8 Introduction

2.3.2. Genome wide association studies (GWASs)

Owing to the sequencing of the by the Human Genome Project in conjunction with the development of high-density genotyping arrays, GWASs have allowed to identify genetic variants with a modest effect on a given disease phenotype (60,61). The GWAS approach is a high-throughput methodology to scan a large number of SNP markers across the entire genome, using powerful statistical and computational methods to analyze associations between a phenotype and different genetic variants through the genome (62). Since GWAS is a hypothesis free approach, it allows the detection of polymorphisms located in genes involved in different pathways that would not have been previously implicated in the phenotype. Once investigators have fully analyzed their GWAS, meta-analyses can be performed to achieve statistical power to uncover additional loci.

More than 50 GWASs and meta-analyses of GWASs of obesity have been carried out since 2006 in children, adolescents and/or adults (Table 5). Insulin induced gene 2 (INSIG2) was the first locus to be reported in a GWAS of obesity; however, replication studies demonstrated very inconsistent results (63). The first locus unequivocally associated with BMI by a GWAS was the fat mass and obesity associated (FTO) gene which was reported by three groups in rapid succession in 2007 (64–66).

Following the discovery of the FTO locus, investigators enhanced GWASs in order to increase the statistical power in the analysis of genetic variants. In this context emerged a multinational collaboration known as the Genetic Investigation of Anthropometric Traits (GIANT) consortium. The first meta-analysis conducted by this consortium confirmed the association of obesity with polymorphisms located in the FTO and identified one new locus near the MC4R gene (67).

In a subsequent meta-analysis of GWAS, the same consortium reported the association of the two above-mentioned loci (FTO and MC4R) and further 6 loci (transmembrane protein 18 (TMEM18), potassium channel tetramerization domain containing 15 (KCTD15), glucosamine- 6-phosphate deaminase 2 (GNPDA2), SH2B adaptor protein 1 (SH2B1), mitochondrial carrier homolog 2 (MTCH2) and neuronal growth regulator 1 (NEGR1)) showing significant association with BMI (68). Four of these loci (TMEM18, NEGR1, KCTD15 and SH2B1) were also observed in a GWAS of the deCODE genetics group that also reported novel loci on BDNF, SEC16 homolog B, endoplasmic reticulum export factor (SEC16B) and ETS variant 5 (ETV5) genes (69).

The third GIANT meta-analysis found 18 new loci associated with obesity near or in: DNaJ heat shock protein family (Hsp40) member C17 (DNAJC17), G protein-coupled receptor class C group

9 Introduction

5 member B (GPRC5B), mitogen-activated protein kinase kinase 5 (MAP2K5), glutaminyl- peptide cycotransferase like (QPCTL), TNNI3 interacting kinase (TNNI3K), solute carrier family 39 member 8 (SLC39A8), 3-hydroxy-3-methylglutaryl-CoA (HMGCR), leucine rich repeat neuronal 6C (LRRN6C), transmembrane protein 160 (TMEM160), fanconi anemia complementation group L (FANCL), cell adhesion molecule 2 (CADM2), protein kinase D1 (PRKD1), LDL receptor related protein 1B (LRP1B), polypyrimidine tract binding protein 2 (PTBP2), mitochondrial translational initiation factor 3 (MTIF3), zinc finger protein 608 (ZNF608), ribosomal protein L27a (RPL27A), and nudix hydrolase 3 (NUDT3) (70). Moreover, in 2013, the GIANT consortium confirmed 32 loci previously associated with obesity and revealed 7 new loci: zinc finger ZZ-type containing 3 (ZZZ3), regulatory associated protein of MTOR complex 1 (RPTOR), adenylate cyclase 9 (ADCY9), G protein subunit alpha transducin 2 (GNAT2), mitochondrial ribosomal protein S33 pseudogene 4 (MRPS33P4), heparin sulfate 6-O- sulfatransferase 3 (HS6ST3) and hepatocyte nuclear factor 4 gamma (HNF4G) (71).

In the largest meta-analysis of GWAS on BMI carried out to date, including over 238,944 subjects in the discovery stage and a further 102,754 subjects in the replication stage, the GIANT consortium identified 97 BMI-associated loci, 56 of which were novel (for example ATP/GTP binding protein like 4 (AGBL4), cell adhesion molecule 1 (CADM1), transcription factor 7 like 2 (TCF7L2), syntaxin binding protein 6 (STXBP6), hypoxia inducible factor 1 alpha subunit inhibitor (HIF1AN), erb-b2 receptor tyrosine kinase 4 (ERB34), fragile histidine triad (FHIT), neuron navigator 1 (NAV1) and RAS p21 protein activator 2 (RASA2)) (72). These 97 loci explained »2.7% of BMI variation.

Although to date most of obesity GWASs and meta-analysis of GWASs have been performed in populations of European ancestry, there are some investigations executed in Asian, Hispanic, African, African American or Native American populations. Results of a large meta-analysis of GWAS using individuals of African origin, found three previously identified loci at SEC16B, GNPDA2 and FTO as well as two new loci at MC4R and polypeptide N- acetylgalactosaminyltransferase 10 (GALNT10) (73). In addition of the previously mentioned loci, when the GIANT consortium was included, Monda et al.(2013) reported three loci; one of them previously identified at adenylate cyclase (ADCY3) gene and two novel loci located in or near kelch like family member 32 (KLHL32) and microRNA 148a-nuclear factor erythroid 2 like 3 (MIR148A-NFE2L3) (73). On the other hand, Wen et al. (2014) carried out a meta-analysis to examine the association of »2.5 million polymorphisms with BMI among Asian ancestry individuals (74). The authors identified 5 novel BMI-associated polymorphisms near the potassium voltage-gated channel subfamily Q member (KNCQ1), aldehyde dehydrogenase 2

10 Introduction family (ALDH2), myosin light chain 2 (MYL2), inter-alpha-trypsin inhibitor chain family member 4 (ITIH4) and 5’-nucleotidase, cytosolic II (NT5C2).

Table 5. Obesity GWASs and meta-analyses of GWASs n (discovery/ Number of SNPs Reference Population replication Obesity phenotype reaching significance study) levels of p< 5x10-8

Herbert et al., 2006 (63) European 694 Obesity -

Dina et al., 2007 (64) European 2,900/ 5,100 Obesity, BMI 1 for obesity

Fox et al., 2007 (75) European 1,341 BMI -

Frayling et al., 2007 (65) European 10,657/ 29,596 BMI -

Hinney et al., 2007 (76) European 929/ 2,269 Extreme obesity -

Scuteri et al., 2007 (66) European, African 4,741/ 3,205 BMI, HC 1 for HC American, Hispanic Liu et al., 2008 (77) European 1,000/ 3,812 BMI -

Loos et al., 2008 (67) European 16,876/ 74,059 BMI 1

Cotsapas et al., 2009 (78) European 3,972 Extreme obesity 1

Lowe et al., 2009 (79) Oceanian 2,906 BMI, weight -

Meyre et al., 2009 (80) European 1,416 Obesity 3

Norris et al., 2009 (81) Hispanic American 219/ 1,190 BMI -

Polasek et al., 2009 (82) European 898 BMI -

Thorleifsson et al., 2009 (69) European 72,598/ 43,615 BMI, weight 21 for BMI, 19 for weight

Willer et al., 2009 (68) European 32,387/ 59,082 BMI 8

Johansson et al., 2010 (83) European 3,925 BMI, weight -

Liu et al., 2010 (84) European 11,536 BMI -

Scherag et al., 2010 (85) European 2,258/ 5,552 Early-onset extreme 4 obesity Speliotes et al., 2010 (70) European 123,865/ BMI 32 125,931 Tönjes et al., 2010 (86) European 948 BMI -

Benjamin et al., 2011 (87) European 9,091/ 2,872 BMI -

Croteau-Chonka et al., 2010 (88) Filipino 1,780 BMI, weight -

Jiao et al., 2011 (89) European 327/ 10,338 Morbidly obesity 2

Malhotra et al., 2011 (90) Pima Indian 1,120/ 2,133 BMI -

Paternoster et al., 2011 (91) European 5,373/ 29,181 BMI 3

Wang et al., 2011 (92) European 1,060/ 1,196 Extreme obesity, 7 BMI, HP, weight Bradfield et al., 2012 (93) European 13,848/ 4,888 Early-onset obesity 7

Comuzzie et al., 2012 (94) Hispanic 815 BMI, weight and 1 for weight z-score change of weight z- change score Dorajoo et al., 2012 (95) South and East Asian 10,482 BMI -

Melka et al., 2012 (96) European 598 BMI 1

Ng et al., 2011 (97) African American 1,715/ 3,274 BMI -

Okada et al., 2012 (98) East Asian 26,620/ 35,625 BMI 7

Wang et al., 2012 (99) European, African 5,218/ 762 BMI - American

11 Introduction

Table 5. Continuation n (discovery, Number of SNPs Reference Population replication Obesity phenotype reaching significance study) levels of p< 5x10-8

Wen et al., 2012 (100) East and South Asian 22,762/ 53,215 BMI 10

Yang et al., 2012 (101) European 133,154/ BMI 2 59,325 Berndt et al., 2013 (71) European 16,068/ 9,791 BMI, Obesity 10 for BMI, 29 for obesity 1, 20 for obesity 2, 1 for obesity 3 and 23 for overweight Graff et al., 2013 (102) European 13,627/ 16,253 BMI 7

Kim et al., 2013 (103) East Asian 756/ 1,301 BMI -

Melén et al., 2013 (104) European 12,359 BMI -

Monda et al., 2013 (73) African American, 37,956/ 27,661 BMI 8 Sub-Saharan African Namjou et al., 2013 (105) European 2,860 BMI -

Plourde et al., 2013 (106) French Canadian 926 BMI -

Wheeler et al., 2013 (107) European 6,889 Severe early-onset 9 obesity Pei et al., 2014 (108) European 20,913/ 7,719 BMI 3

Scannell et al., 2014 (109) South Asian 5,354 BMI, change BMI 1 change BMI over time over time (females) Wen et al., 2014 (74) East Asian 82,438/ 47,352 BMI 5

Yang et al., 2014 (110) East Asian 597/ 2,955 BMI -

Anderson et al., 2015 (111) Australian 361 BMI -

Locke et al., 2015 (72) European, African 23,8944/ BMI 97 American, Hispanic 102,754 McQueen et al., 2015 (112) European, African 1,886 Change BMI over - American, Native time American Warrington et al., 2015 (113) European 9,377/ 3,918 BMI 4

Felix et al., 2016 (114) European 34,744/ 11,313 BMI 15

Hwang et al., 2016 (115) East Asian 484/ 10,390 BMI -

Lee et al., 2016 (116) East Asian 77 BMI -

Ried et al., 2016 (117) European 133,376/ Body shape 31 for AvPC1, 124 for 39,904 AvPC2, 45 for AvPC3, 75 for AvPC4 Salinas et al., 2016 (118) Hispanic, Asian, 5,885/ 10,250 BMI 1 among Hispanic African American, European Justice et al., 2017 (119) European, Asian, 241,258 BMI 58 African Abbreviations: BMI, Body mass index; HC, Hip circumference; AvPC, Average principal component

Visceral fat, in contrast to total body fat, is associated specifically with the development of obesity-related metabolic disorders such as type 2 diabetes or cardiovascular diseases (120). In this context, GWASs and meta-analysis of GWASs have been executed for central obesity related traits: waist circumference and waist-to-hip ratio (Table 6) (121).

The first GWAS that reported a locus associated with central obesity related traits at genome wide significance level was the study by Chambers et al. in 2008 (122). The authors identified two polymorphisms located close the MC4R gene related to waist circumference. In 2009, 2

12 Introduction meta-analyses of GWAS reported at genome wide significance level 3 loci associated with waist circumference (neurexin 3 (NRXN3), transcription factor AP-2 beta (TFAP2B) and methionine sulfoxide reductase A (MSRA)) and 1 locus with waist-to-hip ratio (lysophospholipase like 1 (LYPLAL1)) (123,124).

Since BMI, waist circumference and waist-to-hip ratio are heavily correlated, some of the polymorphisms resulted from GWAS of central obesity related traits are overlapping with that of overall BMI (125). One strategy to capture mechanisms involved in fat distribution more than obesity itself is to adjust waist circumference or waist-to-hip ratio for BMI. In this sense, Heid et al. (2010), Liu et al. (2013), Randall et al. (2013), Shungin et al. (2015) and Wen et al. (2016) reported at genome wide significant level several polymorphisms related to waist circumference or waist-to-hip ratio adjusted for BMI (126–130).

The meta-analysis conducted by Heid et al. (2010) revealed 14 loci associated with waist-to-hip ratio adjusted BMI in European individuals (the polymorphisms located near or in R-spondin 3 (RSPO3), vascular endothelial growth factor (VEGFA) and T-box 15-tryptophanil tRNA synthetase 2 mitochondrial (TBX15-WARS2) genes showed the strongest association) (126). In a population of African ancestry Liu et al. (2012) further identified 2 loci influencing waist circumference adjusted BMI and waist-to-hip ratio adjusted BMI, LIM homeobox 2 (LHX2) and ras responsive element binding protein (RREB1), respectively (127). Results of the study by Randall et al. (2013) in a population of European ancestry showed 4 previously identified loci near growth factor receptor bound protein (GRB14), LYPLAL1, VEGFA and ADAM metallopeptidase with thrombospondin type 1 motif 9 (ADAMTS9); and 3 novel loci near mitogen-activated protein kinase kinase kinase 1 (MAP3K1), hydroxysteroid 17-beta dehydrogenase 4 (HSD17B4) and PPARG (128). In the largest GWAS regarding central obesity related traits, which involved 142,762 individuals in the discovery stage and 67,326 in the replication study of European ancestry, 49 loci, 33 of which were new, were reported to be associated with waist-to-hip ratio adjusted BMI (among the novel loci coiled-coil domain containing 92 (CCDC92), bone morphogenetic protein 2 (BMP2) and leucine, glutamate and lysine rich 1 (LEKR1) showed the strongest association) (129). Additionally, Wen et al. (2016) reported 4 new loci associated with waist circumference after adjustment for BMI (EGF containing fibulin like extracellular matrix protein 1 (EFEMP1), ADAMTS like 3 (ADAMTSL3), canopy FGF signaling regulator 2 (CNPY2) and GNAS complex locus (GNAS)) as well as 2 new loci associated with waist-to-hip ratio after adjustment for BMI (nidogen 2 (NID2) and major histocompatibility complex, class II, DR beta 5 (HLA-DRB5)), in East Asian ancestry individuals (130).

13 Introduction

Table 6. Central obesity GWASs and meta-analyses of GWASs Number of SNPs n Central obesity reaching Reference Population (discovery/ phenotype significance levels replication study) -8 of p< 5x10

Fox et al., 2007 (75) European 1,341 WC - Chambers et al., 2008 (122) European 2,684 WC, WHR 2 for WC Heard-Costa et al., 2009 (123) Caucasian 31,373/ 38,641 WC 1 Lindgren et al., 2009 (124) European 21,397/ 6,021- WC, WHR 2 for WC, 1 for WHR 20,213 Cho et al., 2009 (131) East Asian 8,842/ 16,703 WHR 1 Lowe et al., 2009 (79) Oceanian 2,906 WC - Polasek et al., 2009 (82) Croatia 898 WC - Heid et al., 2010 (126) European 77,167/ 113,636 WHR adj BMI 14 Croteau-Chonka et al., 2011 (88) Filipino 1,792 WC - Kilpeläinen et al., 2011 (132) Whites, Asians, 36,626/ 39,576 WC 1 African American, Hispanic Kim et al., 2013 (103) East Asian 756/ 1,301 WC - Kraja et al., 2011 (133) European 22,161 WC-TG, WC-HDLc 6 for WC-TG, 3 for WC-HDLc Wang et al., 2011 (92) European 1,060 WC, WHR 1 for WC, 1 for WHR Kristiansson et al., 2012 (134) European 10,564 WC 1 Berndt et al., 2013 (71) European 10,077/ 6,250 WHR 4 Liu et al., 2013 (127) African 23,564/ 10,027 WC adj BMI, WHR 1 for WC adj BMI, 1 adj BMI for WHR ajd BMI Randall et al., 2013 (128) European, 73,137/ 74,657 WHR adj BMI 7 European-American descent Shungin et al., 2015 (129) European 142,762/ 67,326 WC, WHR, WHR 19 for WC and WHR adj BMI and 49 for WHR adj BMI Scott et al., 2016 (135) South Asian 10,318/ 1,922 WHR adj BMI - Wen et al., 2016 (130) East Asian 44,609/ 25,553 WC, WHR, WC adj 2 for WC, 4 for WC BMI, WHR adj BMI adj BMI, 2 for WHR adj BMI Justice et al., 2017 (119) European, Asian, 241,258 WC adj BMI, WHR 62 for WC adj BMI, African adj BMI 32 for WHR adj BMI Abbreviations: WC, Waist circumference; WHR, Waist-to-hip ratio; WHR adj BMI, Waist-to-hip ratio adjusted for body mass index; TG, Triglycerides; HDLc, High density lipoprotein cholesterol; WC adj BMI, Waist circumference adjusted for body mass index

In addition, GWASs and meta-analysis of GWASs of more direct measurements of body composition and body fat distribution measured by bioelectrical impedance, computed tomography (CT) or DXA have been published in different populations (Table 7). However, up to date a limit number of studies have reported polymorphisms associated with body composition measurements reaching genome wide significance levels of p< 5x10-8.

In this context, Kilpeläinen et al. (2011) reported 3 loci associated with body fat mass by DXA or bioimpedance among subjects of European and Indian Asian origin (insulin receptor substrate 1 (IRS1), sprouty RTK signaling antagonist 2 (SPRY2) and FTO) (136). In the study by Pei et al. (2014) of European ancestry individuals, 2 loci reached genome-wide significance, one for body fat mass (MC4R) and the other one for body fat mass adjusted for body lean mass (cathepsin S (CTSS)) (108). Lu et al. (2016) in a population of European, Asian and African

14 Introduction

American ancestry individuals, identified 12 loci related to body fat mass, of which eight were previously associated with increased overall adiposity (BMI or body fat mass) (FTO, IRS1, MC4R, TMEM18, SPRY2, translocase of outer mitochondrial membrane 40 (TOMM40), tu translation elongation factor, mitochondrial (TUFM), SEC16B) and four were novel associations with body fat mass (in or near cordon-bleu WH2 repeat protein like 2 (COBLL1), insulin like growth factor 2 mRNA binding protein 1 (IGF2BP1), phospholipase A2 group VI (PLA2G6) and CREB regulate transcription coactivator 1 (CRTC1)) (137). In the last GWAS on body composition traits, 7 loci attained genomic wide significance (138). Of the 7 loci, 2 were associated with subcutaneous adipose tissue (gasdermin B (GSDMB) and FTO), 2 were associated with visceral adipose tissue adjusted for BMI (GRAM domain containing 3 (GRAMD3) and RREB1) and 3 were associated with relative fat distribution (visceral adipose tissue/subcutaneous adipose tissue ratio adjusted for BMI; ubiquitin conjugating enzyme E2 E2 (UBE2E2), LYPLAL1 and lymphocyte antigen (LY86)).

Table 7. Body composition GWASs and meta-analysis of GWASs n Number of SNPs Reference Population (discovery/ Obesity phenotype reaching significance replication study) levels of p< 5x10-8

Liu et al., 2008 (77) Caucasian 1000 Body fat mass by DXA 1 Lowe et al., 2009 (79) Oceanian 2906 Body fat mass by BIA - Norris et al., 2009 (139) Hispanic American 229/ 1,190 VAT, SAT, VAT/SAT - ratio by CT Kilpeläinen et al., 2011 European, Indian 36,626/ 39,576 Body fat mass by DXA 3 (136) Asian or BIA Fox et al., 2012 (140) European, African- 1,057/ 3,158 SAT, VAT, SAT/VAT by 1 for VAT/SAT, 1 for American CT VAT among women Melka et al., 2012 (96) European 598 Body fat mass by BIA 1 Plourde et al., 2013 (141) French Canadian 926 Body fat, SAT, VAT by - CT Pei et al., 2014 (108) European 21,969/ 6,663 Body fat mass, body 1 for body fat mass, fat mass adj lean mass 1 for body fat mass by DXA adj lean mass Tan et al., 2014 (142) Caucasia, Asian, 11,161 Body fat mass by DXA, - African-American, or underwater Hipanic weighing Lu et al., 2016 (137) European, South 100,716 Body fat mass by DXA, 12 and East Asian, bioimpedance or African American underwater weighing

Sung et al., 2016 (143) European, African 2,513/ 2,943 Body fat mass, VAT, 11 for SAT among American SAT, VAT/SAT women Chu et al., 2017 (138) European, African, 18,332 VAT, SAT, VAT/SAT 2 for SAT, 2 for VAT Hispanic, Asian ratio, VAT adj BMI, SAT adj BMI, 3 for adj BMI, VAT/SAT adj VAT/SAT and BMI by CT VAT/SAT adj BMI Abbreviations: DXA, Dual-energy-X-ray absorptiometry; BIA, Bioimpedance; VAT, Visceral adipose tissue; SAT, Subcutaneous adipose tissue; VAT/SAT, Visceral adipose tissue/subcutaneous adipose tissue ratio; CT, Computed tomography; adj, Adjusted; BMI, Body mass index

15 Introduction

2.3.3. Genetic risk-allele scores (GRSs)

Several GWASs have identified a large number of obesity susceptibility polymorphisms. Nevertheless, the effect size for any one SNP on the phenotype is small due to potentially many other genes are regulating the same phenotype and are operating in an interactive fashion (144). In this context, combining information from different obesity loci into genetic risk-allele scores (GRSs) could be a convenient way to summarize risk-associated variations across the genome (145). Generally, the GRSs may explain a considerable proportion of variation in the phenotype, even if one of the SNPs individually does (37).

The simple way to calculate a GRS is by summing the number of accumulated risk alleles associated with the disease (Table 8). Using this approach, Li et al. (2010) created a GRS based on 12 BMI-associated polymorphisms in a sample of 20,341 individuals of European descent (146). Each additional risk allele was associated with increases of 0.15 kg/m2 in BMI, 0.44 kg in body weight and 0.36 cm in waist circumference; and 10.8% and 5.5% increased risk of obesity and overweight, respectively. The GRS explained 0.9% of the phenotypic variance in BMI with an area under the curve (AUC) of 0.57 for prediction of obesity. In a South Asian population, Ahmad et al. (2015) combined a total of 95 polymorphisms in a GRS which was associated with 0.04 kg/m2 higher BMI and an odds ratio for obesity of 1.02 per each BMI-raising allele (147). In this study, the GRS explain 1.54% of BMI variation.

Another approach to calculate a GRS is by effect-size weighting, under which risk alleles are weighted by the effect size reported for each locus in previous GWASs or meta-analysis of GWASs (Table 8). In this sense, Belsky et al. (2013) performed a weighted GRS based on b coefficients derived from the GIANT Consortium or the DeCode BMI GWAS (69,70,145). The authors reported that the 32-locus GRS was significantly associated with BMI and obesity with an AUC of 0.57. Moreover, the GRS by Renström et al. (2009), which compressed 9 polymorphisms, was associated not only with BMI and obesity risk but also with body composition markers. Those subjects in the top quintile of the score showed higher body weight, total fat, gynoid fat and abdominal fat than those in the lowest quintile (148).

16 Introduction

Table 8. Selection of GRSs associated with obesity related traits in observational studies among adults Number Reference Population Outcome of SNPs Simple count GRS

Li et al., 2010 (146) 20,431 European 12 GRS was associated with obesity and overweight risk GRS was associated with BMI, weight and WC Klimentidis et al., 2011 294 African American, 16 GRS was associated with body fat mass adj height2 among (149) Hispanic, European Hispanic American Peterson et al., 2011 2,653 European- 56 GRS was associated with BMI and obesity risk (150) American, 973 African- American Lemas et al., 2013 (151) 1,073 10 Interaction between GRS and n-3 PUFA on BMI, fat mass and WC Rukh et al., 2013 (152) 26,107 13 GRS was associated with BMI, fat mass, fat free mass and energy and fiber intake Ahmad et al., 2015 16,157 South Asian 95 GRS was associated with BMI an obesity risk (147) No interaction between GRS and PA and smoking on BMI Nettleton et al., 2015 63,317 European 32 BMI BMI-GRS and WHR-GRS were associated with BMI and BMI (153) 14 WHR adj WHR Interaction between WHR-GRS and diet score on BMI adj WHR Weighted GRS based on published ßs

Renström et al., 2009 4,923 Swedish 9 GRS was associated with weight, BMI, total adiposity, (148) abdominal adiposity and gynoid adiposity Speliotes et al., 2010 8,120 European 32 GRS was associated with BMI (70) Qi et al., 2012 (154) 12,304 32 GRS was associated with BMI Interaction between GRS and PA and TV watching on BMI Qi et al., 2012 (155) 33,097 32 GRS was associated with BMI Interaction between GRS and sugar-sweetened beverages on BMI Ahmad et al., 2013 111,421 12 GRS was associated with BMI (156) Interactions between GRS and PA levels on BMI Jääskeläinen et al., 2013 459 26 GRS was associated with BMI and WC (157) Mägi et al., 2013 (158) 2,812 European 32 GRS was associated with severe obesity (BMI ≥35 Kg/m2), but not with BMI Martínez-García et al., 2,294 Spanish 6 GRS was associated with BMI 2013 (159) Goumidi et al., 2014 1,578 French 31 GRS was associated with BMI and obesity risk (160) Qi et al., 2014 (161) 9,623 32 GRS was associated with BMI Interaction between GRS and fried food consumption on BMI Locke et al., 2015 (72) 8,164 European 97 GRS was associated with BMI and obesity risk

Simple count GRS and Weighted GRS based on published ßs

Belsky et al., 2013 (145) 10,745 white and 32 GRS was associated with BMI and obesity risk African American Domingue et al., 2014 918 European 31 GRS was associated with BMi and obesity risk (162) Fox et al., 2014 (163) 6,272 32 GRS was associated with BMI Sandholt et al., 2014 3,982 European 32 GRS was associated with BMI at baseline but not with (164) changes in body weight after 5 years Hung et al., 2015 (165) 2,521 European 32 GRS was associated with BMI No interaction between GRS and depression status on BMI Simple count GRS and Weighted GRS based on published ORs and on ORs of the study

Cheung et al., 2010 1,170 Chinese 13 GRSs were associated with obesity risk (166) Abbreviations: GRS, Genetic risk score; BMI, Body mass index; WC, Waist circumference, adj, Adjusted; n-3 PUFA; Omega-3 polyunsaturated fatty acids; PA, Physical activity; WHR, Waist-to-hip ratio

17 Introduction

2.3.4. Nutrigenetics, gene-environment interactions

The genetic variants identified so far only explain a small proportion of heritability of the diseases, indicating huge “missing” heritability (167). It is believed that at least a part of the missing heritability of obesity could be explained through interactions between genetic variants and environmental exposures such as lifestyle and dietary factors (168). At this point the term nutrigenetics, which is the contraction of the words nutrition and genetics, should be introduced. Nutrigenetics is the science that investigates the combined effect of genetic variation and nutrition on health and performance needed for personalized nutrition (169). Gene-environment interactions may reflect a causal mechanism where the genetic variations and environmental exposures contribute to the causation of a phenotype (Figure 2) (Table 9).

BODY’S GENETIC SYSTEM ENVIRONMENTAL INPUTS METABOLIC RESPONSE

Figure 2. Gene-environment interactions

Dietary habits are associated with obesity risk and this association could be modified by genetic background (121,168,170). In this context, significant gene-dietary fat interactions have been reported not only for total fat but also for saturated fatty acids (SFA), monounsaturated fatty acids (MUFA) or different polyunsaturated fatty acids (PUFA) including omega-3 and omega-6 fatty acids. For example, Corella et al. (2011) observed that the effects of total fat, SFA and MUFA intake significantly interacted with rs9939609 FTO genetic variant on BMI (171). In contrast to these results, in a large-scale analysis based on data from 177,330 adults, the strongest obesity polymorphism identified so far did not interact with dietary intakes on BMI on overall population (172). Other genes that have demonstrated gene-diet interactions with dietary fat intake on obesity related traits are apolipoprotein A2 (APOA2) with SFA (173), apolipoprotein B (APOB) with total fat (174), lipoprotein lipase (LPL) with PUFA (175), adiponectin C1Q and collagen domain containing (ADIPOQ) with omega-3 PUFA (176) and ADAM metallopeptidase domain 17 (ADAM17) with omega-6 PUFA (177). In another study

18 Introduction examined whether carbohydrates intake interacted with perilipin (PLIN) and PPARG genetic variants in relation to adiposity measurements (178). The authors observed that when carbohydrate intake was high (≥144 g/day), carriers of the variant allele of the rs894160 PLIN genetic variant exhibited smaller waist and hip circumferences compared with the non-variant allele carriers. In contrast, when complex carbohydrate was low (<144 g/day), variant allele carriers exhibited larger waist circumference.

Physical activity is among the most important lifestyle factors that affect not only the risk of obesity, but also all-cause mortality (179). Although several studies reported that the rs9939609 FTO risk allele is attenuated in individuals who are physically active (180,181), other studies have not been able to replicate such interaction (182,183). In this sense, a large meta- analysis of 218,166 adults confirmed that the association of the rs9939609 FTO variant with BMI and obesity risk was approximately 30% smaller in physically than in inactive individuals (132). In line of this results a recent genome-wide interaction analysis comprising 200,452 adults found that the effect of the rs9941349 FTO variant, which is in strong linkage desequilibirum with rs9939609 polymorphism, was attenuated by 33% in physically active subjects compared to inactive subjects (184). In addition, Klimentidis et al. (2016) found a significant gene-environment interaction between the rs9939609 FTO genetic variant and time spent sitting on BMI among European-American, whereby the association of time spent sitting with BMI was greatest among those carriers of the risk allele (185). However, among Hispanic and African-American individuals the interaction term was in the opposite direction, obtaining conflicting results.

Other lifestyle behaviors may also interact with genetic factors in affecting obesity risk. Ahmad et al. (2015, 2016) in two consecutive studies, in subjects of Pakistani origin of the same cohort; reported significant interactions between genetic variants located near or in PTBP2, huntingtin interacting protein 1 (HIP1), chromosome 6 open reading frame 106 (C6orf106), glutamate ionotropic receptor delta type subunit 1 (GRID1) and uncharacterized LOC285150 (FLJ33534) genes and smoking on BMI (147,186). In another study, Edwards et al. (2012) examined gene-environment interactions with known obesity risk factors including smoking and alcohol consumption, among others (187). The authors reported that smoking and alcohol consumption significantly interacted with 5-hydroxytryptamine (serotonin) receptor 2C (HTR2C), adiponectin receptor 1 (ADIPOR1), insulin receptor (INSR) and insulin like growth factor binding protein 3 (IGFBP3), and with PPARG coactivator 1 alpha (PPARGC1A) and LEPR, respectively, on BMI. Sleep duration has also been assessed as a modifying factor in the association between genetics and obesity related traits. For example, Dashti et al. (2015)

19 Introduction observed an interaction between sleep duration and melatonin receptor 1B (MTNR1B) rs1387153 for BMI, which suggests 0.25 and 0.60 kg/m2 higher BMIs with short (<7 hours) and long (≥9 hours) sleep durations, respectively (188). Other factors that seem to interact with genetic variants influencing adiposity traits are education and socioeconomic status (189,190).

Table 9. Selected gene-diet interactions in observational studies among adults Environmental Major finding Reference Population Genetic factor factor

Smith et al., 2008 Hispanic rs894160, Total fat and CH Complex CH * rs894160 PLIN (178) rs2289487, variant on BMI, WC and HC rs2304795 (PLIN), rs181282 (PPARG) Corella et al., 2011 2,163 European, rs9939609 and Total fat, SFA, PUFA, Total fat, SFA and MUFA (171) Hispanic rs1121980 (FTO) MUFA, CH intake * FTO variants on BMI Kilpeläinen et al., 2011 218,166 Whites, rs9939609 (FTO) PA PA * FTO variant on BMI, (132) Asian, African- overweight and obesity risk, American, Hispanic WC, body fat percentage Lagou et al., 2011 621 European- rs1799983 (NOS3) Socioeconomic Socioeconomic status * NOS3 (190) American, African- status variant on body fat American percentage Corella et al., 2012 1,580 Caucasian rs9939609 (FTO) Education level Education level * FTO variant (189) on BMI Edwards et al., 2012 2,338 African- HTRC2, ADIPOR1, PA, smoking, alcohol, PA, smoking, alcohol, time (187) American, PPARGC1A, PPARA, time spent sitting, spent sitting or time spent Caucasian INSIG2, LEPR, INSR, time spent sleeping sleeping * different genetic IGFBP3, PPARG, variants on BMI among GHRL African-American or Caucasians Qi et al., 2014 (172) 177,330 Whites, rs9939609 (FTO) Energy, total fat, Energy and macronutrient African-American, protein, CH intake not interacted with Asian FTO variant on BMI in overall population Ahmad et al., 2015 16,157 Pakistan CLIP, CADM2, PA, smoking PA * CLIP, CADM2 and (147) GALNT10, PTB2, GALNT10 variants on BMI HIP1, C6orf106, Smoking interacted with GRID1 PTB2, HIP1, C6orf106 and GRID1 variants on BMI Dashti et al., 2015 28,190 European rs1387153 Sleep duration Sleep duration * MTNR1B (188) (MTNR1B) variant on BMI Ahmad et al., 2016 14,131 Pakistani rs140133294 Smoking Smoking * FLJ33534 variant (186) (FLJ33534) on BMI Klimentidis et al., 2016 12,074 European- rs9939609 (FTO) Time spent sitting Time spent sitting * FTO (185) American, African- variant on BMI American, Hispanic Graff et al., 2017 (184) 200,452 European, rs9941349 (FTO) PA PA * FTO variant on BMI other ancestry Abbreviations: *, Interacted with; PLIN, Perilipin; PPARG, Peroxisome proliferator activated receptor gamma; FTO, Fat mass and obesity associated; NOS3, Nitric oxide synthase 3; HTR2C, 5-hydroxytryptamine (serotonin) receptor 2C; ADIPOR1, Adiponectin receptor 1; PPARGC1A, PPARG coactivator 1 alpha; INSIG2, Insulin induced gene 2; LEPR, Leptin receptor; INSR, Insulin receptor; IGFBP3, Insulin like growth factor binding protein 3; GHRL, Ghrelin and obestatin prepropeptide; CLIP, CAP-Gly domain containing linker protein 1; CADM2, Cell adhesion molecule 2; GALNT10, Polypeptide N-acetylgalactosaminyltransferase; PTB2, Polypyrimidine tract-binding protein 2; HIP1, Huntingtin interacting protein 1; C6orf106, Chromosome 6 open reading frame 106; GRID1, Glutamate ionotropic receptor delta type subunit 1; MTNR1B, Melatonin receptor 1B;FLJ33534, Uncharacterized LOC285150; MUFA, Monounsaturated fatty acids; PUFA, Polyunsaturated fatty acids; BMI, Body mass index.

In addition to individual polymorphisms, more studies have focused on examine interactions with GRSs (Table 8). In two studies of American men and women from three prospective cohorts, the Nurses’ Health Study (NHS), the Health Professional’s Follow-up Study (HPFS) and the Women Genome Health Study (WGHS); assessed interactions between sugar-sweetened beverages and fried food consumption, and a GRS of 32 obesity-associated genetic variants and BMI (155,161). In one hand, in the NHS and the HPFS the increases in BMI per 10 risk

20 Introduction alleles were 1.00 for sugar sweetened beverage intake of <1 serving/month, 1.03 for 1-4 servings/month, 1.39 for 2-6 servings/week and 1.77 for ³1 servings/day (p for interaction <0.001) (155). The findings were further replicated in the WGHS cohort. On the other hand, in the combined three cohorts, the differences in BMI per 10 risk alleles were: 1.1, 1.6 and 2.2 for fried food consumption less than once, one to three times, and four or more times per week, respectively (p for interaction <0.001) (161).

According to GRS-lifestyle interactions Qi et al. (2012) and Ahmad et al. (2013) assessed interactions between GRSs and physical activity and TV watching (154,156). The weighted 32 BMI-associated loci GRS calculated by Qi et al. (2012) showed an interaction with hours of TV watching on BMI (154). An increment of 10 points in the GRS was associated with 0.8, 0.8, 1.4, 1.5 and 3.4 kg/m2 higher BMI across the 5 categories of TV watching; 0-1, 2-5, 6-20, 21-40 and >40 hours per week, respectively (p for interaction 0.001). The study by Ahmad et al. (2013) found a statistically significant interaction between an obesity GRS based on 12 polymorphisms previously related to obesity and physical activity on BMI (p for interaction 0.015) (156).

Up to date, the investigation of gene-diet interactions on obesity are based on genetic variants identified in candidate gene approach or GWASs. Meanwhile, genome-wide approaches have been successfully applied in the detection of gene-environment interactions in other disorders such as hypertension, there is little evidence of genome-wide analysis on gene-environment interactions in relation to obesity (191).

3. GENETICS OF BODY WEIGHT REGULATION

Obesity has reached epidemic proportions becoming a major global health challenge (1). Therefore, a number of strategies have been investigated in order to induce a negative energy balance and body weight loss, such as reduction of energy intake, and increase in physical activity, behavioral approaches and pharmacological or surgical treatments (15,192,193). However, individual responses to body weight loss interventions vary widely and several studies have aimed to identify psychological, behavioral and personal predictors of this variability (194–196).

In this sense, genetic factors have been described to be associated with adiposity and body weight control, since there are genes involved in the regulation of energy expenditure, appetite, thermogenesis, adipogenesis, insulin resistance and lipid metabolism (Figure 3) (197).

21 Introduction

Figure 3. Karyogram depicting loci that have been associated with body weight loss in response to a nutritional intervention. Source: Goni et al. J Nutr, 2016 (197)

3.1. Genes involved in energy expenditure

Some authors have observed a greater than expected reduction in the metabolic rate during caloric restriction, suggesting that the body attempts to compensate for reduced energy intake to conserve energy stores (198). Although the reason for this phenomenon is still not completely understood, genetic make-up could partially explain it. In fact, several studies have reported the association between genes involved in energy expenditure and body weight regulation such as ADRB2 and ADRB3, which are involved in the regulation of thermogenesis and lipolysis in white and brown adipose tissue (199–203); and UCP1, UCP2 and UCP3, which are implicated in thermoregulation, regulation of ATP synthesis, control of oxidative stress, and fatty acids and glucose oxidation (204–208).

3.2. Genes involved in appetite control and food intake

Much attention has been focused on the role of the hypothalamic leptin-melanocortin system and other systems related to food intake and body weight regulation (209). In this sense,

22 Introduction polymorphisms located in LEP and LEPR, both involved in satiety regulation, energy expenditure, immune and inflammatory responses, lipid and carbohydrate metabolism, and nutrient intestinal absorption, among other processes (210); have been reported to modify the response to weight loss treatments (211–213).

The hypothalamus plays a pivotal role in energy homeostasis by controlling appetite and satiety. Thus, genetic variants in several of the genes that participate in appetite control such as MC4R and BDNF have been related to significant differences in the weight loss response to different hypocaloric diets (214,215).

It is well known that serotonin and dopamine, 2 neurotransmitters controlling different central and peripheral functions, play an important role in the regulation of appetite and body weight (216). In this context, an association between polymorphisms located in the HTR2C and dopamine receptor D2 (DRD2) genes, with weight loss after different behavioral interventions has been reported (217,218).

The endocannabinoid system has received much attention as a mediator of food intake and palatability regulation (219). In this sense, 2 recent studies reported the association between the fatty acid amide hydrolase (FAAH) and the cannabinoid receptor 2 (CNR2) genetic variants with weight loss (220,221).

3.3. Genes involved in adipogenesis and lipid metabolism

Genes implicated in the regulation of adipocyte growth and differentiation can be considered as candidate genes for body weight regulation. Thus, PPARG which encodes a regulatory protein of adipocyte differentiation, has been associated with body weight loss (222). Two polymorphisms located in the TCF7L2 gene, which plays an important role in adiposity through several mechanisms (223), have been associated not only with body weight loss, but also with the regulation of body composition, because they were related to changes in waist circumference, fat mass, fat-free mass, or visceral fat (224–227).

Although not yet investigated thoroughly, functional interactions between adipose tissue and components of the lipoprotein transport system have been suggested (228). In this context, an apolipoprotein A5 (APOA5) genetic variant, which has an important role in regulating plasma triglycerides concentrations, has been reported to be associated with body weight regulation (229). Other genetic variants located in or near some genes related to lipid metabolism, such as acyl-CoA synthetase long-chain family member 5 (ACSL5), PLIN1, fatty acid binding protein

23 Introduction

2, intestinal (FABP2), and Niemann-Pick Disease, type C1 (NPC1), have been related to resistance to body weight loss (230–236).

3.4. Other polymorphisms of interest

Although the relationship between the FTO gene and obesity has unequivocally been confirmed across different ethnic and demographic populations, the association with body weight loss is not clear. To date, 3 polymorphisms located in FTO have been associated with the response to calorie restriction (222,237–239). Nevertheless, a recent meta-analysis of 11 clinical trials have not confirmed such association (240).

Other studies on the etiology of obesity have evaluated the role of common polymorphisms located in genes involved in clock systems, because circadian disruptions may contribute to different metabolic-related traits (241). In regard to body weight regulation, the clock circadian regulator (CLOCK) gene and the aralkylamine N-acetyltransferase (AANAT) gene have been related to the response to different weight loss interventions (222,242).

3.5. Gene-diet interactions on body weight regulation

In addition, it has been reported whether a specific diet or its macronutrient content might differentially affect weight loss depending on the presence of certain SNP. In the Preventing Overweight Using Novel Dietary Strategies (POUNDS Lost) trial, which is a randomized, controlled 2-years weight loss trial using 4 diets that differed in macronutrient proportions, have performed a series of analyses on gene-diet interactions and changes in adiposity markers (243). For example, participants with the rs2943641 CC genotype of the IRS1 gene showed greater weight loss than T allele carriers in the high-carbohydrate diet group (244). Another study indicated that individuals carrying the C allele of the fibroblast growth factor 21 (FGF21) rs4253449 polymorphism might benefit more in changes in central adiposity and body composition measurements when undertaking a high-carbohydrate/low-fat diet (245).

The Nutrient-Gene Interactions in Human Obesity: implications for dietary guidelines (NUGENOB) study, a 10-weeks randomized clinical trial in which overweight or obese participants were randomly assigned to one of the two weight loss diets, also reported gene- diet interactions on body weight regulation (246). Grau et al. (2010) observed that homozygous for the T allele of the rs7903146 TCF7L2 genetic variant who were assigned to a high-fat diet showed smaller weight loss, waist circumference and fat free mass than those of

24 Introduction the low-fat diet group (224). Moreover Stocks et al. (2012) reported that dietary fat intake interacted with genetic variants located in or near TFAP2B, catenin beta like 1 (CTNNBL1) and NPC intracellular cholesterol transporter 1 (NPC1) on changes in weight or waist circumference (236).

The Diet, Obesity and Genes (DiOGenes) study is a randomized, controlled 6-month dietary intervention study to determine the effects of dietary protein and glycemic index on weight regain and metabolic risk factors after an 8-week intensive weight loss period (247). In the study by Larsen et al. (2012), in which 768 SNPs in presumed nutrient-sensitive genes were examined, the genetic variants in ghrelin and obestatin prepropeptide (GHRL), cholecystokinin (CCK), MLX interacting protein like (MLXIPL) and LEPR showed interactions with dietary protein intake on weight regain; and the genetic variants in peroxisome proliferator activated receptor delta (PPARD), fatty acid binding protein 1 (FABP1), lipin 1 (LPIN1) and plasminogen activator, urokinase receptor (PLAUR) showed interactions with dietary protein on fat mass regain (248).

Table 10. Gene-diet interactions in randomized intervention studies for weight loss Reference Population Intervention Genetic factor Major finding

Grau et al., 2009 618 overweight/ 10-w weight loss, rs9939609 (FTO) Dietary fat * FTO variant on (249) obese 2 diets changes in energy expenditure Grau et al., 2010 622 overweight/ 10-w weight loss, rs7903146 (TCF7L2) Dietary fat * TCF7L2 variant on (224) obese 2 diets changes in weight, WC and body composition Garaulet et al., 1,465 overweight/ 28-w weight loss, rs1801282 (PPARG) Dietary fat * PPARG variant on 2011 (250) obese MedDiet changes in weight Qi et al., 2011 738 overweight/ 6-m weight-loss, 4 rs2943641 (IRS1) Diet intervention * IRS1 variant (244) obese diets on changes in weight Larsen et al., 2012 742 overweight/ 6-m weight-loss GHRL, LEPR, CCK, Dietary protein * different (248) obese maintenance, 5 MLXIPL, PPARD, FABP1, genetic variants on changes in diets LPIN1 and PLAUR weight regain or fat mass regain Heni et al., 2012 304 overweight/ 9-m weight-loss, 2 rs7903146 (TCF7L2) Dietary fiber * TCF7L2 variant (227) obese diets on changes in weight Mattei et al., 2012 591 overweight/ 2-y weight-loss, 4 rs7903146 and Dietary fat * TCF7L2 variants on (226) obese diets rs12255372 (TCF7L2) changes in weight and body composition Qi et al., 2012 662 overweight/ 6-m weight-loss, 4 rs2287019 (GIPR) Dietary fat * GIPR variant on (251) obese diets changes in weight Stocks et al., 2012 642 overweight/ 10-w weight- loss, rs987237 (TFAP2B) Dietary fat * TFAP2B, CTNNBL1 (236) obese 2 diets rs6013029 (CTNNBL1) and NCP1 on changes in weight and rs1805081 (NPC1) or WC Zhang et al., 2012 742 overweight/ 2-y weight-loss, 4 rs1558902 (FTO) Dietary protein * FTO variant on (237) obese diets changes in weight and body composition Xu et al., 2013 658 overweight/ 2-y weight-loss, 4 rs1440581 (PPM1K) Dietary fat * PPM1K variant on (252) obese diets changes in weight Huang et al., 2014 609 overweight/ 6-m weight-loss, 4 rs9939609 (FTO) Dietary protein * FTO on (253) obese diets changes in appetite and food cravings Mirzaei et al., 2014 722 overweight/ 2-y weight-loss, 4 rs10830963 (MTNR1B), Dietary fat * MTNR1B and CRY2 (254) obese diets rs11605924 (CRY2) variants on changes in RQ Lin et al., 2015 723 overweight/ 2-y weight-loss, 4 rs16147 (NPY) Dietary fat * NPY variant on (255) obese diets changes in weight and body composition Heianza et al., 715 overweight/ 2-y weight-loss, 4 rs838147 (FGF21) Dietary carbohydrate * FGF21 2016 (245) obese diets variant on changes in weight, WC and body composition

25 Introduction

Table 10. Continuation Reference Population Intervention Genetic factor Major finding Larsen et al., 2016 11,048 2-y weight-loss rs4253449 (CEBPB) Dietary fat, MUFA and PUFA * (256) overweight/ obese maintenance, 5 CEBPB variant on being a body diets weight gainer Loria-Kohen et al., 179 overweight/ 16-w weight-loss rs3749474 (CLOCK) Dietary fat * CLOCK variant on 2016 (257) obese changes in BMI Abbreviations: *, Interacted with; MedDiet, Mediterranean diet; FTO, Fat mass and obesity associated; TCF7L2, Transcription factor 7 like 2; PPARG, Peroxisome proliferator activated receptor gamma; IRS1, Insulin receptor substrate 1; GHRL, Ghrelin and obestatin prepropeptide; LEPR, Leptin receptor; CCK, Cholecystokinin; MLXIPL, MLX interacting protein like; PPARD, Peroxisome proliferator activated receptor delta; FABP1, Fatty acid binding protein 1; LPIN1, Lipin 1; PLAUR, Plasminogen activator urokinase receptor; GIPR, Gastric inhibitory polypeptide receptor; TFAP2B, Transcription factor AP-2 beta; CTNNBL1, Catenin beta like 1; NPC1, NPC intracellular cholesterol transporter 1; PPM1K, Protein phosphatase Ma2+/Mn2+ dependent 1K; MTNR1B, Melatonin receptor 1B; CRY2, Cryptochrome circadian clock 2; NPY, Neuropeptide Y; FGF21, Fibroblast growth factor 21; CEBPB, CCAAT/enhancer binding protein beta; CLOCK, Clock circadian regulator; WC, Waist circumference; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; BMI, body mass index; RQ, Respiratory quoeficient

4. GENETICS OF OBESITY METABOLIC RELATED TRAITS

Obesity is associated with and may contribute to the development of many metabolic diseases including type 2 diabetes, hypertension, hyperlipidemia, cardiovascular disease and cancer (258). Unfortunately, the increased prevalence of obesity worldwide has led to an increase in the prevalence of its related comorbidities (258). A better understanding of the causes of the obesity related traits is essential to mitigate the immense medical and economic impact of them. Interestingly, as obesity, such metabolic comorbidities are in part genetically influenced.

4.1. Hyperlipidemia

Hyperlipidemia is the most common form of dyslipidemia, characterized by abnormal elevated levels of any or all lipids or lipoproteins in the blood. The key plasma lipids are total cholesterol and triglycerides which need to be transported in the plasma associated with various lipoprotein particles since are nonpolar lipids (259). There are five major classes of plasma lipoproteins: chylomicrons, very-low-density lipoproteins (VLDL), intermediate-density lipoproteins (IDL), low-density lipoproteins (LDL) and high-density lipoproteins (HDL). The hyperlipidemias are classified according to the Fredrickson classification which was later adopted by the WHO (Table 11) (260).

Each of these phenotypes include different diseases caused by specific genetic abnormalities in which generally environmental factors are involved (primary hyperlipidemias) or acquired when resulting from another underlying disorder that leads to alterations in plasma lipids and/or lipoproteins (261). Primary hyperlipidemia includes hypercholesterolemia, hypertriglyceridemia and combined hyperlipidemia (Table 12). On the other hand, the secondary hyperlipidemia can be caused by diabetes, use of certain drugs, hypothyroidism, kidney failure, nephrotic syndrome and the diet, among other factors (261).

26 Introduction

Table 11. Fredrickson classification of hyperlipidemias Phenotype Increased lipoprotein Plasma cholesterol Plasma triglycerides Type I Chylomicrons Normal or ­ ­­­­ Type II a LDL ­­ Normal b LDL and VLDL ­­ ­­ Type III IDL ­­ ­­­­ Type IV VLDL Normal or ­ ­­ Type V Chylomicrons and VLDL ¹ ­­ ­­­­ Abbreviations: LDL, Low-density lipoprotein; VLDL, Very-low-density lipoprotein; IDL, Intermediate- density lipoprotein Adapted from: Mataix. Nutrición y Alimentación Humana, 2015 (262)

Table 12. Primary hyperlipidemias Plasma Plasma Fredrickson Increased Altered Hyperlipidemia Frequency cholesterol triglycerides classification lipoprotein genes (mg/dL) (mg/dL) Hypercholesterolemia Monogenic familial Homo >300 1/500 IIa LDL - LDLR hypercholesterolemia Hete 300/600 Polygenic familial 1/20 IIa LDL <300 - Unknown hypercholesterolemia Defected ApoB 100 Variable IIa LDL 275-500 - APOB hypercholesterolemia Hypertriglyceridemia Familial Unusual I Chylomicrons - 1,200/1,500 LPL, APOC2 hyperchylomicronemia Endogenous familial 1/300 IV VLDL - 200/500 Unknown hypertriglyceridemia Hypertriglyceridemia of Chylomicrons 1/500 V >300 >1,000 Unknown mixed origin and VLDL Combined hyperlipidemia Combined familial 1/100 IIa, IIb, IV LDL and VLDL <400 ­ Unknown hyperlipidemia Familial 1/5,000 III b-VLDL 300/1,000 300/1,000 APOE, LIPC dysbetalipoproteinemia Abbreviations: LDL, Low-density lipoprotein; VLDL, Very-low-density lipoprotein; Homo, Homocygote; Hete, Heterocygote; LDLR, Low density lipoprotein receptor; APOB, Apolipoprotein B; LPL, Lipoprotein lipase; APOC2, Apolipoprotein C2; APOE, Apolipoprotein E; LIPC, Lipase C hepatic type Adapted from: Mataix. Nutrición y Alimentación Humana, 2015 (262)

Abnormal metabolism of lipids and lipoproteins is considered an intermediate phenotype of cardiovascular diseases (263,264). In fact, the improvement of lipid profiles has been the therapeutic target of such diseases (265). Moreover, hyperlipidemias have been associated with other chronic diseases such as diabetes and certain type of cancers (266–268).

In 2008, the global prevalence of raised cholesterol among adults was 39% (37% among males and 40% among females) (269). Globally, between 1980 and 2008, mean total cholesterol changed little, decreasing by less than 0.1 mmol/L per decade in both men and women (270). High total cholesterol is estimated to cause 2.0 million (1.6 million to 2.5 million) of deaths annually and 1.6% (1.3% to 2.0%) of disability–adjusted live years (271). Analysis of 30-year national trends in serum lipid levels demonstrates improvements in total cholesterol and LDL

27 Introduction cholesterol (265). However, this trend may in part be explained by the increase in the use of lipid-lowering drugs.

4.1.1. Genetics of hyperlipidemia

The investigation of genetic variants associated with plasma lipid and lipoprotein levels began with twin studies, family studies, linkage analyses and candidate gene association studies. Thus far, most of the well-characterized disorders of lipoprotein metabolism are monogenic, familial disorders with extreme phenotypes (Table 12). For example, familial hypercholesterolemia is a caused by mutations in the low-density lipoprotein receptor (LDLR). Heterozygous familial hypercholesterolemia occurs with a prevalence of 1/200 to 1/500, and it has been estimated that there are 20 millions of patients worldwide (272,273). The genetic disorder is characterized by high plasma LDL cholesterol levels and early-onset cardiovascular disease (274).

Regarding candidate gene studies, several lipid genes such as upstream regulatory factor 1 (USF1), ww domain containing oxidoreductase (WWOX), cholesteryl ester transfer protein (CETP), apolipoprotein E (APOE) or lipase c hepatic type (LIPC) have been identified (275–279). Although, the results of most of them have never been replicated, meta-analyses of candidate gene studies supported some of the reported associations between lipid levels and candidate lipid genes (280). Moreover, a large candidate gene association resource which encompassed more than 40,000 individuals, support the association of HDL cholesterol levels with CETP, LIPC and LPL, LDL cholesterol with proprotein convertase subtilisin/kexin type 9 (PCSK9), and serum triglycerides with LPL and apolipoprotein A5 (APOA5) (281).

Discovery of common variants associated with the trait has been substantially facilitated by the application of GWAS through which hundreds of novel genomic loci associated not only with lipid levels but also with cardiovascular and other diseases have been identified (282,283). The first GWAS of circulating lipids, which was published in 2007, found a novel genetic variant located in an intronic region of the glucokinase regulatory protein (GCKR) gene related to serum triglycerides (284). Later several GWAS and meta-analysis of GWAS have been carried out in different ethnic populations (Table 13). In one of the last GWAS a total of 93 loci were associated with 4 lipid traits (HDL cholesterol, LDL cholesterol, total cholesterol and triglycerides) (285). Of them, 4 loci presented a well-established function in lipid genetics (cadherin EGF LAG seven-pass G-type receptor 2 (CELSR2), GCKR, LIPC and APOE) and 2 loci

28 Introduction could be considered candidate missense mutations with predicted damaging function (CD300 molecule like family member g (CD300LG), transmembrane 6 superfamily member 2 (TM6SF2)).

Table 13. Selection of lipid traits GWASs and meta-analyses of GWASs n Lipid metabolism SNPs reaching significance Reference Population (discovery/ -8 trait levels of p< 5x10 replication study)

Saxena et al., 2007 (284) European 5,217 TG, LDL-c, HDL-c, 1 for TG, 1 for LDL-c and 1 apoAI, apoB for HDL-c Kathiresan et al., 2008 (286) European 2,578/ 18,544 TG, LDL-c, HDL-c 8 for TG, 8 for LDL–c and 5 for HDL-c Willer et al., 2008 (287) European 8,684/ 9,741 TG, LDL-c, HDL-c 10 for TG, 9 for LDL–c and 14 for HDL-c Kooner et al., 2008 (288) European, Asian, 2,011/ 10,536 TG, HDL-c 5 for TG and 4 for HDL-c Hispanic Kamatani et al., 2010 (289) Asian 14,402 TG, HDL-c 4 for TG and 3 for HDL-c

Teslovich et al., 2010 (290) European 100,184 TG, TC, LDL-c, 32 for TG, 52 for TC, 36 for HDL-c LDL-c and 47 for HDL-c Kim et al., 2011 (291) Asian 12,545/ 30,395 HDL-c 2

Coram et al., 2013 (292) African-American, 7,917/ 7,138 TG, LDL-c, HDL-c 4 for TG, 6 for LDL-c and 9 Hispanic for HDL-c Willer et al., 2013 (293) European 94,595/ 93,982 TG, TC, LDL-c, 39 for TG, 74 for TC, 58 for HDL-c LDL-c and 71 for HDL-c Ko et al., 2014 (294) Hispanic 3,701/ 6,017 TG, TC, HDL-c 3 for TG, 2 for TC and 6 for HDL-c Surakka et al., 2015 (285) European 58,358 TG, TC, LDL-c, 93 loci associated with HDL-c lipid traits. 4 for TG, 3 for TC, 3 for LDL-c and 2 fo HDL-c new loci van Leeuwen et al., 2016 (295) European, 49,549/ 73,519 TG, TC, LDL-c, 3 for TG, 2 for HDL-c new African-American, HDL-c loci Asian

Abbreviations: TG, Triglycerides; LDL-c, Low density lipoprotein cholesterol; HDL-c, High density lipoprotein cholesterol; TC, Total cholesterol

Not surprisingly, many of the identified polymorphisms affect multiple lipid fractions. In this sense, a GWAS in nearly 187,000 individuals, which identified 157 independent loci associated with lipid levels at genome wide significant level, found that 36 loci were associated with both total cholesterol and LDL cholesterol, and 4 loci showed associations with total cholesterol, HDL cholesterol, LDL cholesterol and triglycerides, for example (Figure 4) (293).

29 Introduction

Figure 4. Venn diagram that illustrates the number of loci that show association with different lipid traits. The number of loci primarily associated with only one trait is listed in parentheses after the trait name, and locus names are listed below. Loci that show association with two or more traits are shown in the appropriate segment. Source: Global Lipids Genetics Consortium. Nat Genet, 2013 (293)

4.1.2. Gene-environment interactions on hyperlipidemia

A difference in the effect size of a genetic variant in individuals differing in an environmental exposure suggests the presence of gene-environment interactions. Thus far, several studies on gene-environment interactions in determining lipid metabolism traits have been carried out (Table 14). In one of the largest gene-environment study, which involved 27,756 individuals of the CHARGE consortium, found an LPL-PUFA intake interaction in determining triglyceride levels (296). Besides, the MLXIPL rs3812316 genotype was reported to modify the association between Mediterranean diet adherence and serum triglycerides (297). Although the diet has been the lifestyle behavior most studied, there are other factors such as smoking, alcohol intake and physical activity that also interact with genetic susceptibility in determining plasma lipids (298,299). Initially most results on gene-environment interactions came from observational studies. However, in recent years new data are coming from dietary intervention studies. For example, in the DiOGenes study reported that LPIN1 rs4315495 genetic variant

30 Introduction interacted with dietary protein intake to modify changes in triglycerides in response to a 6- month weight loss maintenance period (300).

Table 14. Selected gene-environment interactions in observational studies and clinical trials among adults Environmental Reference Population Genetic factor Major finding factor

Volcik et al., 2008 10,134 rs1800206, rs6008259, n-3 and n-6 Dietary n-3 PUFA * PPARA rs3892755 (301) Whites, 3,480 rs3892755 (PPARA) PUFA variant on TC and LDLc in African African- American American Richardson et al., 3,605 rs884164, rs1609717, n-3 and n-6 Dietary n-3 and n-6 PUFA * PLIN4 2011 (302) European rs7250947, rs8887, PUFA rs884164 variant on TG. n-6 PUFA * rs8102428, rs892158, PLIN4 rs884164 variant on HDLc. rs11673616 (PLIN4) Hellstrand et al., 4,6335 rs174546 (FADS1) n-6 PUFA Dietary n-6 PUFA * FADS1 variant on 2012 (303) European LDL-c Brahe et al., 2013 841 European 240 SNPs in 42 candidate 6-m weight Dietary protein * LPIN1 rs4315495 (300) genes maintenance, variant on changes in TG 2 diets Grammer et al., 3,263 APOE (Alleles E2, E3 and Smoking Smoking * APOE variant on LDL-c, 2013 (298) European E4) oxidized LDLc and lipoprotein- associated phospholipase A2 Richardson et al., 27,756 rs13702 (LPL) PUFA Dietary PUFA * LPL variant on TG 2013 (296) Corella et al., 2014 7,187 rs13702 (LPL) MUFA, MUFA and unsaturated fat intake * (304) European unsaturated LPL variant on baseline and 3-year fat changes TG Ortega-Azorín et 7166 rs3812316 (MLXIPL) MedDiet MedDiet * MLXIPL variant on TG al., 2014 (297) European Qi et al., 2014 732 American rs37644261 (CETP) 2-y weight Dietary CH * CETP variant on changes (305) and African- loss, 4 diets in triglycerides and HDLc. The results American were replicated in an independent cohort Son et al., 2015 1,193 Asian rs662799 (APOA5), Alcohol, Alcohol, smoking and PA * APOA5 (299) rs769450 (APOE) smoking, PA and APOE variants on TG Abbreviations: n-3, Omega-3; n-6, Omega-6; PUFA, Polyunsaturated fatty acids; MUFA, Monounsaturated fatty acids; MedDiet, Mediterranean diet; PA, Physical activity; *, Interacted with; TC, Total cholesterol; LDLc, Low density lipoprotein cholesterol; TG, Triglycerides; HDLc, High density lipoprotein cholesterol; PPARA, Peroxisome proliferator activated receptor alpha; PLIN4, Perilipin 4; FADS1, Fatty acid desaturase 1; LPIN1, Lipin 1; APOE, Apolipoprotein E; LPL, Lipoprotein lipase; MLXILP, MLX interacting protein like; CETP, Cholesteryl ester transfer protein; APOA5, Apolipoprotein A5

4.2. Type 2 diabetes

Type 2 diabetes is a metabolic disorder characterized by chronic as a result of a relative insulin deficiency caused by pancreatic β-cell dysfunction and insulin resistance (IR) in target organs (306). Insulin is a hormone secreted by the b cells of the pancreatic islets, whose main function is the regulation of carbohydrate and protein metabolism in the body (307). Thus, IR is defined as the inability of cells, primarily muscle, liver and fat tissue to respond adequately to normal levels of insulin (308).

Either the fasting plasma glucose, the 2-h plasma glucose value after a 75 g oral glucose tolerance test (OGTT) and glycated hemoglobin (Hb1Ac) are equally appropriate criteria for type 2 diabetes diagnostic (309). The current American Diabetes Association (ADA) and WHO recommendations for the diagnostic criteria for diabetes are summarized in Table 15. In addition, has been recognized an intermediate group of individuals whose glucose levels do

31 Introduction not meet criteria for diabetes but are too high to be considered normal. These people have been defined as having impaired fasting glucose (IFG) or impaired glucose tolerance (IGT) (Table 15). Individuals with IFG or IGT have been referred as pre-diabetic.

Table 15. The WHO and ADA classifications of type 2 diabetes WHO (310) ADA (311) Type 2 diabetes ³7.0 mmol/L (126 mg/dL) ³7.0 mmol/L (126 mg/dL) Fasting plasma glucose or or ³11.1 mmol/L (200 mg/dL) ³11.1 mmol/L (200 mg/dL) 2-h plasma glucose* or or ³6.5% HbA1c ³6.5% or ³11.1 mmol/L (200 mg/dL) and Random plasma glucose - classic symptoms of hyperglycemia or hyperglycemic crisis Prediabetes IGT <7.0 mmol/L (126 mg/dL) Fasting plasma glucose - and ³7.8 and <11.1 mmol/L (140 and 200 7.8 to 11.0 mmol/L (140 to 200 2-h plasma glucose* mg/dL) mg/dL) Prediabetes IFG 6.1 to 6.9 mmol/L (110 to 125 mg/dL) Fasting plasma glucose 5.6 to 6.9 mmol/L (100-125 mg/dL) and (if measured) 2-h plasma glucose* <7.8 mmol/L (140 mg/dL) - Prediabetes HbA1c - 5.7% - 6.4% * Venous plasma glucose 2 hours after ingestion of 75 g oral glucose load Abbreviations: IGT, Impaired glucose tolerance; IFG, Impaired fasting glucose; HbA1c; Glycated hemoglobin; WHO, World Health Organization; ADA, American Diabetes Association

The risk of type 2 diabetes is determined by an interplay of genetic, metabolic and environmental factors. Ethnicity, family history of type 2 diabetes, and previous combine with aging, overweight and obesity, unhealthy diet (high intake of total fat, SFA and sugar, and inadequate consumption of dietary fiber), physical inactivity and smoking increase the risk of the disorder (310,312). The complications of type 2 diabetes, if it is not well controlled, are diabetic retinopathy, renal disease, diabetic neuropathy, cardiovascular disease (coronary artery disease, cerebrovascular disease, peripheral vascular disease and myocardial ischemia), and foot ulcers that could lead in lower-extremity amputations (310).

The global rise in obesity, unhealthy lifestyles an ageing population, have quadrupled the incidence and prevalence of diabetes from 108 million in 1980 to 422 million in 2014, being most (85-95%) cases of type 2 diabetes (312). In 2012, 1.5 million deaths worldwide were caused by diabetes, being the eighth leading cause of death (310). In addition, the International Diabetes Federation (IDF) has estimated that the prevalence of type 2 diabetes by 2040 will be 642 million (313).

32 Introduction

4.2.1. Genetics of type 2 diabetes

Twin and familial studies support the principle of inherited genetic susceptibility as an important risk factor for type 2 diabetes. In a meta-analysis of 8 twin cohorts, which involved 34,166 same-sex twin pairs, the heritability of type 2 diabetes was estimated at 72% (314). Furthermore, offspring of a parent with type 2 diabetes present a 40% lifetime risk of developing type 2 diabetes, which increases to 70% when both parents are diabetic (315). In this context, experts of the ADA, the Juvenile Diabetes Research Foundation (JDRF), the European Association for the Study of Diabetes and the American Association of Clinical Endocrinologist, in the Differentiation of diabetes by pathophysiology, natural history and prognosis research symposium (October, 2015) agreed that both, genetic and environmental factors can result in the progressive loss of b-cell mass and/or function; and concluded that the identification of individualized therapies for diabetes will require better characterization of the many pathways involved in the metabolic disorder (316).

Studies of single gene disorders that exhibit the features of diabetes have been a useful model to get the first insights into the genetics of the trait. There are two different classes of monogenic diabetes where hyperglycemia is either due to defects in insulin secretion, decrease in b-cell mass or both: maturity onset diabetes of the young (MODY) and neonatal diabetes mellitus (NDM) (Table 16). MODY is a clinical subgroup of familial diabetes that showed autosomal dominant inheritance, positive familial history, early age onset, absence of auto-immune antibodies and IR (317). On the other hand, NDM is characterized by onset hyperglycemia in the first few weeks of life and can be either transient NDM or permanent NDM (318). In addition to these categories, there are some monogenic forms of IR such as primary lipodystrophic syndromes and insulin receptor defects leading to Donahue syndrome, Rabson Mendenhall syndrome or Type-A insulin resistance (315).

Further evidence comes from candidate gene studies which have focused their search on genes that encode proteins in the pathways of glucose-induced insulin secretion from the pancreatic β-cells, peripheral insulin-induced glucose uptake in muscle and fat, and insulin regulation of liver gluconeogenic pathways (319). In fact, genes involved in pancreatic β-cell function like ATP binding cassette subfamily C member 8 (ABCC8), potassium voltage-gated channel subfamily J member 11 (KCNJ11), solute carrier family 2 (facilitated glucose transporter) member 2 (SLC2A2), hepatocyte nuclear factor 4 alpha (HNF4A), insulin (INS), and genes influencing insulin action such as PPARG, INSR, phosphoinositide-3-kinase regulatory (PIK3R1), IRS1 and SOS Rac/Ras guanine nucleotide exchange factor 1 (SOS1) were among the

33 Introduction initially identified candidate genes that significantly were associated with type 2 diabetes risk (320).

Table 16. Selection of different forms of monogenic diabetes and insulin resistance Disease Gene Chr location Phenotype Maturity-onset diabetes of the young (MODY) MODY-1 HNF4A 20q13.12 Familial, early-onset diabetes Mild and long-lasting stable hyperglycemia, MODY-2 GCK 7p15.3-p15.1 mutations can cause PNDM Normoglycemic in childhood, develop progressive b- MODY-3 HNF1A 12q24.2 cell dysfunction MODY-4 PDX1 13q12. Pancreatic agenesis Neonatal diabetes mellitus (NDM) Permanent NDM (PNDM) CISD2 4q24 Sensorineural hearing loss, optic atrophy or neuropathy, and defective platelet aggregation FOXP3 Xp.11.23 Immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome GLIS3 9p24.2 Congenital hypothyroidism, glaucoma, liver fibrosis and cystic kidney disease NEUROG3 10q21.3 Diabetes and chronic intractable malabsorptive diarrhea starting soon after birth Temporary NDM (TNDM) ZPF57 17 Intrauterine growth restriction SLC2A2 3q26.1-q26.2 Hepatomegaly, proximal tubular nephropathy, hypergalactosemia Monogenic forms of insulin resistance Very low adiponectin levels, acanthosisnigricans and CGL type 1 AGPAT2 12q14.1 hypertrophic cardiomyopathy FPLD-type 3 PPARG 3p25 Excess abdominal fat and hypertension Loss of subcutaneous adipose tissue primarily FLPD-type 4 PLIN1 15q26 affecting lower limbs Donahue syndrome, Ronson- Acanthosisnigricans, extreme hyperinsulinemia but INSR 19p13.3-p13.2 Mendenhall syndrome, Type-A IR normal lipid profile, preserved adiponectin levels Abbreviations: HNF4A, Hepatocyte nuclear factor 4 alpha; GCK, Glucokinase; HNF1A, HNF1 homeobox A; PDX1, Pancratic and duodenal homeobox 1; CISD2, CDGSH iron sulfur domain 2; FOXP3, Forkhead box P3; GLIS3, GLIS family zinc finger 3; NEUROG3, Neurogenin 3; ZPF57, Zinc finger protein 57; SLC2A2, Solute carrier family 2 (facilitated glucose transporter) member 2; AGPAT2, Acylglycerol-3-phosphate O-acyltransferase 2; PPARG, Peroxisome proliferator activated receptor gamma; PLIN1, Perilipin 1; INSR, Insulin receptor; Chr, Chromosome; PNDM, Permanent neonatal diabetes mellitus Adapted from: Tallapragada DSP et al. Frontiers in Genetics, 2015 (315)

In addition, to uncover novel loci contributing to the pathogenesis of the trait, several GWASs have been carried out in different ethnic populations (Figure 5). The first wave of GWAS was conducted among Europeans and found 12 type 2 diabetes susceptibility loci, of which 8 have been replicated across multiple ethnic groups: TCF7L2, solute carrier family 30 member 8 (SLC30A8), hematopoietically expressed homeobox (HHEX), CDK5 regulatory subunit associated protein 1 like 1 (CDKAL1), insulin like growth factor 2 mRNA binding protein 2 (IGF2BP2), cyclin dependent kinase inhibitor 2A/2B (CDKN2A/B), PPARG and KCNJ11 (321–328). Meanwhile the first wave of GWAS investigated type 2 diabetes as a dichotomous phenotype in cases and controls, the second wave of GWAS identified loci associated with continuous glycemic traits such as fasting glucose, fasting insulin, 2-h plasma glucose after OGTT, HbA1c and proinsulin. For example, in one of the last GWAS of the Meta-Analysis of Glucose and Insulin-related traits Consortium (MAGIC), grouped the identified loci into five clusters (329). The first cluster was

34 Introduction associated with insulin sensitivity and including genetic variants located near or in PPARG, kruppel like factor 14 (KLF14), IRS1 and GCKR genes. The second cluster, which included MTNR1B and glucokinase (GCK) variants, was related to reduce insulin secretion and fasting hyperglycemia. The third cluster contained the ArfGAP with RhoGAP domain Ankyrin repeat and PH domain (ARAP1) gene and was characterized by defects in insulin processing. TCF7L2, SLC30A8, HHEX, CDKAL1 and CDKN2A/2B constituted the fourth cluster characterized by loci influencing insulin processing and secretion. The last cluster was performed by 20 risk loci with no clear association to continuous glycemic traits.

Figure 5. GWAS on type 2 diabetes and glycemic traits. Source: Flannick et al. Nature Reviews, 2016 (330)

4.2.2. Gene-environment interactions on type 2 diabetes

Such findings strongly support the concept that genes play a central role in the determination of glucose metabolism traits and consequently in the pathogenesis of type 2 diabetes. However, most of the previous genetic discovery studies did not consider the potential modification effect of environmental factors, since type 2 diabetes is a complex disorder in which genetics and the environment interacts to predispose to the disease. Thus far, studies on gene-environment interactions for diabetes-related traits have been performed in observational studies and clinical trials (Table 17). A recent systematic review reported the effect of 8 significant interactions between SNPs and macronutrients on type 2 diabetes risk: 2 polymorphisms in the TCF7L2 with dietary fiber, another TCF7L2 variant with glycemic load, 1

35 Introduction polymorphism in gastric inhibitory polypeptide receptor (GIPR) with total fat and carbohydrate intake, 1 polymorphism in caveolin 2 (CAV2) with total fat and SFAs, and 1 SNP in peptidase D (PEPD) with erythrocyte phospholipid omega-3 PUFA (331). Moreover, it has been reported the interaction between associated-type 2 diabetes variants and micronutrients, specific foods as well as dietary patterns (332–335). Kanoni et al. (2011) in a 5-cohort meta-analysis found that total dietary zinc intake (food and supplement sources) modify the association between SLC30A8 genetic variant and fasting glucose levels (332). On the other hand, the European Prospective Investigation into Cancer (EPIC)-InterAct consortium reported an interaction between TCF7L2 rs12255372 variant and coffee intake on type 2 diabetes risk (334). Regarding, the study of gene-environment interactions in clinical trials, the POUNDS Lost trial found significant interactions between IRS1 and protein phosphatase Mg2+/Mn2+ dependent 1K (PPM1K) genetic variants with carbohydrate and fat intake, respectively, on changes in insulin and insulin resistance (244,252).

Table 17. Selected gene-diet interactions in observational studies and clinical trials among adults Environmental Reference Population Genetic factor Major finding factor

Cornelis et al., 3,055 rs12255372 (TCF7L2) GL, GI, CH, cereal GL * TCF7L2 variant on T2D risk 2009 (336) American fiber Fisher et al., 2009 27,548 rs7903146 (TCF7L2) Whole grain Whole grain intake * TCF7L2 variant (333) European on T2D risk Fisher et al., 2011 2,862 63 SNPs Fat, SFA, MUFA, Dietary fat and SFA intake * CAV2 (337) PUFA rs2270188 variant on T2D risk Kanoni et al., 2011 34,333 20 SNPs Total dietary zinc Total dietary zinc intake * SLC30A8 (332) European intake variant on fasting glucose levels Qi et al., 2011 738 American rs2943641 (IRS1) 2-y weight loss, 4 Dietary CH * IRS1 variant on changes (244) and African- diets in insulin and HOMA-IR American Hindy et al., 2012 24,799 rs7903146 (TCF7L2) CH, fat, protein, Dietary fiber intake * TCF7L2 variant (338) European fiber on T2D incidence Sonestedt et al., 24,480 rs10423928 (GIPR) CH, fat, protein, Dietary CH and fat intake * GIPR 2012 (339) European fiber, sucrose variant on T2D risk Corella et al., 2013 7,018 rs7903146 (TCF7L2) MedDiet MedDiet * TCF7L2 on fasting glucose (335) European levels Wirström et al., 5,477 rs7903146, rs4506565 Cereal fiber Cereal fiber intake * TCF7L2 variant 2013 (340) European (TCF7L2) on T2D risk Xu et al., 2013 658 American rs1440581 (PPM1K) 6-m weight loss, 4 Dietary fat * PPM1K variant on (252) and African- diets changes in insulin and HOMA-IR American Zheng et al., 2015 915 Asian 9 SNPs Circulating Circulating erythrocyte membrane (341) erythrocyte phospholipid n-3 PUFA * PEPD membrane rs3786897 variant on T2D risk phospholipid n-3 PUFA The InterAct 19,121 9 incretin-related SNPs Whey containing Coffee intake * TCF7L2 rs12255372 Consortium 2016 European diary, cereal fiber, variant on T2D risk (334) coffee, olive oil Abbreviations: SNP, Single nucleotide polymorphism; *, Interacted with; GL, Glycemic load; GI, Glycemic index; CH, Carbohydrates; MedDiet; Mediterranean diet; T2D, Type 2 diabetes; HOMA-IR, Homeostasis model assessment of insulin resistance; TCF7L2, Transcriptor factor 7 like 2; IRS1, Insulin receptor substrate 1; GIPR, Gastric inhibitory polypeptide receptor; PPM1K, Protein phosphatase Mg2+/Mn2+ dependent 1K; CAV2, Caveolin 2; SLC30A8, Solute carrier family 30 member 8; PEPD, Peptidase D

36 Introduction

4.3. Hypertension

Hypertension, also known as high or raised blood pressure, is a condition in which the blood vessels have persistently raised pressure. Blood pressure is measured in millimeters of mercury (mmHg) and is recorded as two numbers, the systolic and diastolic pressures. The systolic blood pressure (SBP) is the maximum pressure in blood vessels and happens when the heart muscle is contracting. Whereas the diastolic blood pressure (DBP) is the minimum pressure in blood vessels and happens when the heart muscle is relaxing.

In most national and international guidelines hypertension is defined as a SBP ≥ 140 mmHg and/or DBP ≥ 90mmHg (342). In addition, hypertension is sub-classified: The 7th report of the Joint National Committee on prevention, detection, evaluation and treatment of high blood pressure (JNC7) distinguished two stages of hypertension, whereas the European Society of Hypertension (ESH) and the European Society of Cardiology (ESC) in their guideline for hypertension treatment differs between 3 stages of hypertension and isolated systolic hypertension (343,344). Bellow the hypertensive range, the JNC7 proposed normal blood pressure as a SBP < 120 mmHg and a DBP < 80 mmHg. In contrast, the ESH and the ESC differed between optimal (SBP <120 and DBP <80) and normal (120-129 and/or 80-87) blood pressure levels (Table 18).

High blood pressure is classified in primary or essential hypertension, and secondary hypertension according to the cause. Primary hypertension is due to nonspecific lifestyle and genetic factors. Lifestyle features include consumption of food containing too much salt and fat, not eating enough fruit and vegetables, smoking, alcohol, lack of exercise and poor stress management (345). Secondary hypertension includes the monogenic forms of hypertension (e.g. glucorticoid-remediable aldosteronism, syndrome of apparent mineralocorticoid excess, Liddle syndrome, Gitelman and Bartter syndromes) and other diseases that lead to hypertension being kidney disease the most common secondary cause of high blood pressure (346). Other causes include endocrine disorders (e.g. Cushing’s syndrome, hyperthyroidism, hypothyroidism, acromegaly) or the use of certain drugs (e.g. antidepressants, glucocorticoids, sex hormones) (347,348).

37 Introduction

Table 18. The JNC7 and ESH-ESC classifications of blood pressure levels JNC7 (343) ESH-ESC (344) Category SBP (mmHg) DBP (mmHg) SBP (mmHg) DBP (mmHg) Optimal - - <120 and <80 Normal <120 and <80 120-129 and/or 80-87 Prehypertension 120-139 or 80-89 130-139 and/or 85-89 Stage 1 hypertension 140-159 or 90-89 140-159 and/or 90-99 Stage 2 hypertension ≥160 or ≥100 160-179 and/or 100-109 Stage 3 hypertension - - ≥180 and/or ≥100 Isolated systolic hypertension - - ≥140 and <90 Abbreviations: JNC7, The 7th report of the Joint National Committee; ESH-ESC, European Society of Hypertension and European Society of Cardiology; SBP, Systolic blood pressure; DBP, Diastolic blood pressure

Hypertension is a major independent risk factor for cardiovascular diseases for all age, race or sex groups, since it increases the risk of coronary artery disease and stroke among other cardiovascular diseases. Further complications of hypertension include cognitive impairment, dementia, hypertensive retinopathy and hypertensive nephropathy (345).

The WHO estimated that in 2008 worldwide 40% of adults age 25 and above have been diagnosed with hypertension (349). The number of people with high blood pressure increased from 594 million in 1975 to 1.13 billion in 2015 (350). The prevalence of hypertension is lowest in high-income countries than in low- and middle-income countries (351). This fact could be confounded by differences in socioeconomic status and the weak health systems of low- and middle-income countries.

According to the latest data of the Global Burden of Disease project, raised blood pressure (SBP > 115 mmHg) continues to be the highest single contributor to the global burden of disease and to global mortality, leading to 9.4 million deaths each year (271). The effect is largely mediated through coronary artery disease and stroke.

4.3.1. Genetics of hypertension

More than 60 years ago, Dr. Page proposed the Mosaic Theory of Hypertension, which brought together interactions among genetics, environment, adaptive, neural, mechanical and hormonal disorders as the basis of high blood pressure (352). Since that, several strategies have been used to elucidate the genetic basis of essential hypertension.

Identification of genes responsible for monogenic hypertension has provided valuable insights into the genomic mechanisms and biological pathways underlying blood pressure regulation. Molecular genetic studies have now recognized mutations in 8 genes that cause Mendelian forms of hypertension (346). These genes are involved in the same physiological pathway, in

38 Introduction the renal-sodium handling (Table 19). A family history of severe hypertension, especially with early onset suggests the presence of a monogenic form of hypertension.

Table 19. Selection of monogenic forms of hypertension Syndrome Mutation Glucocorticoid remediable Duplication of genes encoding aldosterone synthase and 11b-hydroxylase aldosteronism Apparent mineralocorticoid excess Mutations in the gene encoding 11b-hydroxylase Hypertension exacerbated by pregnancy Mutations in the ligand-binding domain of the mineralocorticoid receptor type 2 Mutation in WNK1 and 4 encoding genes. Mutations in at least one of three genes (Gordon’s syndrome) mapped to 1q31-42, 12p-13 and 17p11-q21 Hypertension with brachydactyly Mutations mapped to 12p11.2-12.2 Peroxisome proliferator-activated Missense mutation receptor gamma Liddle’s syndrome Mutations in the SCNN1b- or g-subunit Mutation on RET gene on chromosome 10, VHL gene on chromosome 13, NF1 gene Pheochromocytoma syndromes on chromosome 17 Abbreviations: WNK1, WNK lysine deficient protein kinase 1; SCNN1, Sodium channel epithelial 1; RET, Ret proto-oncogene; VHL, von Hippel- Lindau tumor suppressor; NF1, Neurofibromin 1

Although there have been significant insights into hypertension physiology gained from the study of rare Mendelian disorders impacting on blood pressure, it must be recognized that these mutations are generally quite rare and account for a very small fraction of the variation in blood pressure in the general population. Thus, efforts to identify genes predisposing to hypertension in the general population to date have focused on candidate gene studies and GWASs.

Candidate gene studies have reported common variants in genes related to the renin- angiotensin-aldosterone system, the same pathway implicated in Mendelian forms of hypertension. Among them, it is well known the biochemical or physiological function for blood pressure regulation of the angiotensinogen (AGT), renin (REN), angiotensin-converting enzyme (ACE) and angiotensin II receptor type 1 (AGTR1) genes (353–355). As has been previously mentioned candidate gene studies are hampered by inconsistent findings. However, some of the more recent candidate gene studies have successfully identify reproducible genetic associations in independent samples. Such successes are likely the result of the employment of large sample sized as Takeuchi et al. (2012) applied (356). The authors in a large-scale study confirmed that two candidate variants located in ACE and cytochrome P450 family 11 subfamily B member 2 (CYP11B2) genes were associated with blood pressure traits. Moreover, other candidate gene studies have taken advantage of high-throughput genotyping technology to identify gene variants related to the trait of interest. In this sense, Ganesh et al. (2013) using a cardiovascular gene-centric array, examining ~ 50,000 SNPs from ~ 2,100 candidate genes for cardiovascular phenotypes, identified two novel loci associated with blood pressure (357).

39 Introduction

In contrast to GWASs for other cardiovascular related phenotypes, early GWAS failed to identified any associations with blood pressure at a level of genome-wide significance (358,359). For example, in the Wellcome Trust Case Control Consortium, 24 independent association signals were identified for 6 diseases, with the exception of hypertension (358). The failure of early GWASs created an impetus for the formation of consortia with the purpose of conducting meta-analyses of GWASs in large samples capable of detecting the modest effects of blood pressure genetic variants. In this context, in 2009, two consortia, CHARGE and Global Blood Pressure Genetics (Global BPgen), reported findings of their large-scale GWAS meta-analyses (360,361). Since the 2009 publications, additional large blood pressure GWASs meta-analyses have been conducted in different consortiums including the International Consortium of Blood Pressure (ICBP) (362–364), the Exome chip Blood Pressure Consortium (365,366), the Asian Genetic Epidemiology Network (AGEN) (367,368) and the Candidate Gene Association Resource (CARe) (357). The GWAS by Hoffman et al. (2016) has uncovered the largest number of genetic loci associated with blood pressure measurement through meta- analysis, contributing 316 newly discovered loci (369). In total, around 40 GWASs have identified more than 700 loci robustly associated with hypertension or blood pressure traits (191) (Table 20).

Table 20. Selection of blood pressure traits GWASs and meta-analyses of GWASs n SNPs reaching Blood pressure Reference Population (discovery/ significance levels of trait -8 replication study) p< 5x10

Levy et al., 2009 (360) European 29,136/ 34,433 HT, SBP, DBP 1 for HT, 4 for SBP and 6 for DBP Newton-Cheh et al., 2009 European, South Asian 34,433/ 113,236 HT, SBP, DBP 8 for HT, 3 for SBP and (361) 5 for DBP Ehret et al., 2011 (362) European 69,365/ 133,361 HT, SBP, DBP 11 for HT, 24 for SBP and 26 for DBP Wain et al., 2011 (363) European 74,064/ 48,607 MAP 2

Kato et al., 2015 (370) European, Asian 99,994/ 220,257 SBP, DBP, MAP 3 for SBP, 2 for DBP and 3 for MAP Ehret et al. 2016 (371) European 201,529/ 140,886 SBP, DBP 54 for SBP and 53 for DBP Hoffman et al., 2017 (369) European, African 99,785/ 221,477 SBP, DBP 102 for SBP and 120 for American, Asian, Hispanic DBP Liu et al., 2016 (365) European, African, 146,56/ 180,726 HT, SBP, DBP 4 for HT, 12 for SBP and Hispanic 12 for DBP Surendran et al. 2016 (366) European, Asian 192,763/ 155,063 HT, SBP, DBP 4 for HT, 13 for SBP and 16 for DBP Warren et al. 2017 (372) European 140,886/ 190,070 SBP, DBP 23 for SBP and 40 for DBP Abbreviations: HT, Hypertension; SBP, Systolic blood pressure; DBP, Diastolic blood pressure; MAP, Mean arterial pressure

40 Introduction

4.3.2. Gene-environment interactions on hypertension

In the study by Hoffman et al. (2016), all the associated loci explained 2.9% and 2.5% in SBP and DBP, respectively (369). However, the heritability of blood pressure has been found to range from about 30-60% in pedigree data to up to 70% in twin studies (373). The missing heritability of blood pressure could be attributable to gene-environment interactions since blood pressure is known to be modulated by a variety of lifestyle factors including diet, exercise, smoking, age as well as obesity (364,374–377) (Table 21). In contrast to the gene- environment interaction studies in obesity, which are based on candidate genetic variants, there is data from GWAS examining how genes interact with environmental factors to influence blood pressure traits (191). For example, in the Framingham Heart Study have been carried out 4 genome-wide analysis on gene-environment interactions (378–381). In the first study, which involved 2.5 million of polymorphisms, Simino et al. (2013) reported an interaction between alcohol intake and solute carrier family 16 member 9 (SLC16A9) genetic variant on SBP (378). Later, Basson et al. (2014) applied a genome wide approach including 487,998 polymorphisms to analyze the interaction between genes and education level (379). The authors showed that education level interacted with pleiotrophin (PTN) and TOX high mobility group box family member 2 (TOX2) genetic variants on SBP and DBP, respectively. Moreover, Sung et al. (2015) and Basson et al. (2015) found that alcohol intake modifies the effect of LDL receptor related protein 2 (LRP2), MDS1 and EVI1 complex locus (MECOM), Trafficking protein particle complex 9 (TRAPPC9), collectin subfamily member 10-mal T-cell differentiation protein 2 (COLEC-10-MAL2), cytochrome b5 type B (CYB5B), MN1 proto- oncogene transcriptional regulator (MN1), SAM and SH3 domain containing 1 (SASH1) and kelch like family member 6/24 (KLHL6/KLHL24) genetic variants on SBP (380,381). Other factors that seem to interact with genetic variants influencing blood pressure traits include BMI and consumption of PUFA and sodium (382–384).

Table 21. Selected gene-diet interactions in observational studies among adults Environmental Reference Population Genetic factor Major finding factor

Taylor et al., 2010 868 parents, 436 SNPs parents cohort, BMI Among parents BMI * CHEK2, CD36, (382) 322 off-spring, 95 SNPs off-spring cohort ADD2 and IL15 variants on DBP. African Among off-spring subjects BMI * ALK American and CAPN13 on SBP and DBP, OTOF on SBP, and LTBP1, SRD5A2 and RASGRP3 on DBP Chang et al., 2012 1,134 Asian rs2238152, rs671, Alcohol Alcohol intake * ALDH2 rs2238152 (385) rs2158029 (ALDH2) variant on HT progress Simino et al., 2013 6,882 2.5 million SNPs Alcohol Alcohol intake * SLC16A9 variant on (378) Caucasian SBP Basson et al., 2014 3,836 487,998 SNPs Education Education level * PTN and TOX2 (379) Caucasian level variants on SBP and DBP, respectively Basson et al., 2015 6,710 2,144,020 SNPs Smoking Smoking status * SASH1 and (381) Caucasian KLHL6/KLHL24 variants on SBP

41 Introduction

Table 21. Continuation Environmental Reference Population Genetic factor Major finding factor Sung et al., 2015 6,889 2,485,435 SNPs Smoking Smoking status * LRP2, MECOM, (380) Caucasian TRAPPC9, COLEC10-MAL2, CYB5B and MN1 variants on SBP Tagetti et al., 2015 3,550 rs2108622 (CYP4F2), PUFA Omega-3 FA and alpha-linoleic acid (383) rs1126742 (CYP4A11), intake * CYP4F2 variant on DBP rs890293 (CYP2J2), change over time rs41507953 and rs751141 (EPHX2), rs174547 (FADS1) Li et al., 2016 2,651 Chinese 2.05 million SNPs Sodium Sodium intake * UST variant on SBP, (384) CLGN variant on SBP, DBP and MAP and LOC105369882 variant on DBP Taylor et al., 2016 2,510 African- 761,050 SNPs Smoking Smoking * NEDD8 variant on SBP (386) American Abbreviations: SNP, Single nucleotide polymorphism; *, Interacted with; ALDH2, Aldehyde dehydrogenase 2 family;CYP4F2; Cytochrome P450 family 4 subfamily F member 2; CYP4A11, Cytochrome P450 family 4 subfamily A member 11; CYP2J2, Cytochrome P450 family 2 subfamily J member 2; EPHX2, Epoxide hydrolase 2; FADS1, Fatty acid desaturase 1; CHEK2, Checkpoint kinase 2; CD36, CD36 molecule; ADD2, Adducin 2; IL15, Interleukin 15; ALK, ALK receptor tyrosine kinase; CAPN13, Calpain 13; OTOF, Otoferlin; LTBP1, Latent transforming growth factor beta binding protein 1; SRD5A2, Steroid 5 alpha-reductase 2; RASGRP3, RAS guanyl releasing protein 3; SLC16A9, Solute carrier family 16 member 9; PTN, Pleiotrophin; TOX2, TOX high mobility group box family member 2; LRP2, LDL receptor related protein 2; MECOM, MDS1 and EVI1 complex locus; TRAPPC9, Trafficking protein particle complex 9; COLEC10, Collectin subfamily member 10; MAL2, Mal T-cell differentiation protein 2; CYB5B, Cytochrome b5 type B; MN1, MN1 proto-oncogene transcriptional regulator; SASH1, SAM and SH3 domain containing 1; KLHL, Kelch like family member; UST, Uronyl 2-sulfotransferase; CLGN, Calmegin; NEDD8, Neuronal precursor cell expressed developmentally down-regulated 8; BMI, Body mass index; PUFA, Polyunsaturated fatty acids; SBP, Systolic blood pressure; DBP, Diastolic blood pressure; HT, Hypertension; FA, Fatty acids; MAP, Mean arterial pressure

42

HYPOTHESIS AND OBJECTIVES

Hypothesis and Objectives

1. HYPOTHESIS

Based on the available evidence, on the one hand we proposed that different genetic variants or a genetic risk score influence obesity and its related comorbidities susceptibility or the response to a weight loss diet. On the other hand, focusing on gene-environment interactions we hypothesized that dietary factors may modify the association between the genotype and the studied phenotype.

2. GENERAL OBJECTIVE

The general aim of this work was to study the association between different genetic variants and a genetic risk score with obesity related features such as adiposity, blood pressure, and lipid and glucose metabolism traits, as well as the effect of dietary factors on such relationship.

3. SPECIFIC OBJECTIVES

The specific objectives of the present research were:

1. To search for potential polymorphisms associated with obesity and body weight regulation and to identify other metabolic biomarkers such as epigenetics and gut microbiota related to the obesity phenotype (Chapter 1, Chapter 2).

2. To evaluate the validaty of a food groups frequency questionnaire based on an exchange system for assess dietary habits in a short manner (Chapter 3).

3. To test the association between a multi-trait genetic risk score and body fatness and to identify phenotype and genotype variables related to BMI status (Chapter 4, Chapter 5).

4. To examine the effect of genetic variants on changes in body fatness and body composition measurements in response to different weigh-loss diets varying in macronutrient distribution (Chapter 6, Chapter 7, Chapter 8).

5. To evaluate the potential association between different polymorphisms and obesity related traits (blood pressure, glucose metabolism and lipid metabolism), and to investigate the possible influence of the diet on that association (Chapter 9, Chapter 10, Chapter 11).

45

SUBJECTS AND METHODS

Subjects and methods

In order to study the possible genetic factors underlying obesity and its related disorders as well as potential gene-diet interactions different populations were used in the present work (Figure 6): the Nutrigenetic Service Ns cohort, the OBEKIT study, the POUNDS Lost trial and the NUGENOB study.

Nutrigenetic Service Ns cohort Chapters 4, 5, 7 and 11

§ 23 polymorphisms related to the susceptibility to nutritional diseases (i.e. obesity, hypertension, lipid metabolism) study § Anthropometric and body composition measurements, and blood pressure Transversal n=1050

Diet (30% fat, 20% protein, 50% CH)

§ MTNR1B rs10830963, FTO rs9939609 and MC4R study Intervention period rs17782313 variants Intervention § Anthropometric measurements 3-6 weeks n=167

OBEKIT study Chapter 6

Low-fat diet (22% fat, 18% protein, 60% CH)

Moderately high-protein diet (30% fat, 30% protein, 40% CH)

§ ADCY3 rs10182181 variant Intervention period Maintenance period § Anthropometric and body composition measurements (DXA scan) n=107 16 weeks 10 months

Experimental design POUNDS Lost trial Chapters 8 and 9

Low-fat diet (20% fat, 15% or 25% protein, 65% or 55% CH)

High-fat diet (40% fat, 15% or 25% protein, 45% or 35% CH) § MTNR1B rs10830963 variant § Anthropometric and body Intervention period composition measurements (DXA scan) § Triglycerides, cholesterol, LDL 6 months 2 years n=722 cholesterol, HDL cholesterol

NUGENOB study Chapter 10

Low-fat diet (20-25% fat, 15% protein, 60-65% CH)

High-fat diet (40-45% fat, 15% protein, 40-45% CH)

§ PPM1K rs1440581 variant Intervention period § Fasting plasma glucose, fasting plasma insulin, HOMA-IR and HOMA-B n=757 10 weeks

Figure 6. Experimental design. Abbreviations: CH, Carbohydrates; DXA, Dual-energy-X-ray absorptiometry; POUNDS, Preventing overweight using novel dietary strategies; LDL, Low density lipoprotein; HDL, High density lipoprotein; NUGENOB, Nutrient-gene interactions in human obesity: implications for dietary guidelines; HOMA-IR, Homeostasis model assessment of insulin resistance; Homeostasis model assessment of b-cell function; MTNR1B, Melatonin receptor 1B; FTO, Fat mass and obesity associated; MC4R, Melanocortin receptor 4; ADCY3, Adenylate cyclase 3; PPM1K, Protein phosphatase Ma2+/Mn2+ dependent 1K

In addition, to asses the validity of a food groups frequency questionnaire (FGFQ), based on 19 food groups, we used a specific population.

49 Subjects and methods

1. VALIDATION STUDY OF A FOOD GROUPS FREQUENCY QUESTIONNAIRE (FGFQ)

The assessment of intake and eating habits become increasingly important to relate them to the risk of disease. In this sense, food frequency questionnaires (FFQ) are a common dietary tool used in both clinical practice and nutritional epidemiological studies. In order to guarantee the quality of their measurements it is essential that the FFQ should be validated (387). There are different psychometric characteristics that can be evaluated: reliability, validity, sensitivity and feasibility (387). Validity is the characteristic most studied and is based on assess the degree to which the questionnaire agrees with a gold standard (388). In the present study we evaluate the validity of a FGFQ using a 7-day weighed food record as gold standard.

The study protocol was approved by the Research Ethics Committee of the University of Navarra (ref. 110/2014). The study was performed in accordance to the ethical guidelines of the Declaration of Helsinki (389). All participants provided written informed consent after they received an information sheet and additional verbal explanation of the protocol.

1.1. Study population

The study was conducted from September 2014 through June 2015 at the Centre for Nutrition Research of the University of Navarra. A total of 60 subjects were recruited and enrolled in the study. The sample size was calculated to obtain a correlation coefficient of 0.7 with a lower 95% confidence limit of 0.5. The inclusion and exclusion criteria were as follows (Table 22).

Table 22. Inclusion and exclusion criteria of the validation study Inclusion criteria Adults: 18-70 y.o. BMI: ³18.5 kg/m2 and <40 kg/m2 Exclusion criteria BMI <18.5 or ³40 kg/m2 Modification of the diet in the last year by diagnosis of one or more pathologies, surgical intervention, pregnancy, breastfeeding and weight loss Eating disorders Life expectancy < 1 year Some type of cognitive impairment and / psychic Anti-psychotic or anti-depressant drugs prescription Subjects in which poor collaboration or, in the investigator's opinion, have difficulty following the procedures of the study is foreseen Lack of commitment (at the discretion of the investigator) with the intervention, suspected non-compliance, or real difficulties to follow the development of the study Abbreviations: BMI, Body mass index

50 Subjects and methods

1.2. Data collection

1.2.1. Anthropometric and body composition measurements

Weight and height were measured using a digital scale (OMRON BF400) and a portable stadiometer (Leicester TANITA), respectively. The measurements were collected wearing light clothes and no shoes by a trained nutritionist following standardized procedures.

1.2.2. Food groups frequency questionnaire (FGFQ)

In the FGFQ basic foods were classified into 19 food groups: whole dairy products, half-fat dairy products and fat-free dairy products, eggs, fat meat and sausages, lean meat, white fish and shellfish, blue fish, vegetables, fruits, nuts, legumes, olive oil, other fats and oils, refined grains, whole grains, pastries and confectionary industry, sugars, water and alcohol drinks (Figure 7). Each food group comprised several foods based on a food exchange system (390). In other words, the foods on each group have about the same amount of calories, carbohydrate, protein and fat. For example, the group of whole dairy products (in average 67.6 kcal, 3.9 g carbohydrate, 4.2 g protein and 4.0 g fat) included 125 ml of milk, 50 g of milk powder, 125 g of different types of yogurt and 50 g of fresh cheese. Subjects were asked to report how often (daily, weekly, monthly or never) they had consumed a choice of each food group during the previous year. Total energy and macronutrient intake were calculated using Spanish food composition tables (391–393).

Figure 7. Food groups frequency questionnaire (FGFQ) to validate

51 Subjects and methods

1.2.3. 7-day weighed food record

A 7-day weighed food record was used as reference method to check the validity of the FGFQ (388). The volunteers should note the weight of each of the foods and beverages (including water) consumed over 7 consecutive days, 5 week days and 2 weekend days. The participants were instructed to weight all the food and beverages they consumed and all of them received instructions orally and in writing to complete the weighed food record. In addition, volunteers were encouraged to maintain their habitual diet.

Each of the record sheets was categorized into 6 meals (breakfast, morning snack, lunch, afternoon snack, dinner and other snacks). The volunteers should specify the food or ingredients, oil or fat used, cooking technology, whether the weight was raw or cooked (mainly for pasta, rice and legume), or the type of food according to fat (fat or fat-free products), fiber (whole or refined), or sugar contents, among others. In the case of pre-cooked foods, the brand was required and if possible, the product label should be attached. When the meal was made outside house, it was requested to specify the home measure (i.e. a spoonful of olive oil, two slices of cheese). The calibration of the 7-day weighed food record was performed using the software DIAL (Alce Ingeniería) following validated criteria (394).

2. NUTRIGENETIC SERVICE NS COHORT

The Nutrigenetic Service NS is a genetic tool developed by CINFA laboratories in order to provide personalized nutritional advice based on the genetic background of each individual. The designed nutrigenetic test comprised a total of 23 polymorphisms previously related to the development of nutritional diseases (obesity, hypertension, lipid metabolism disorders, type 2 diabetes, cardiovascular disease, folic acid metabolism, deficit of vitamin D, lactose intolerance and osteoporosis).

Seekers of the nutrigenetic service were specifically asked if they would be willing to take part anonymously in the research study. After ensuring that participants had understood the information, only those who provided written informed consent for participation were enrolled. The survey was in accordance with the principles of the Declaration of Helsinki and patient data were codified to guarantee anonymity accuracy (389). The Research Ethics Committee of the University of Navarra gave confirmation of fulfillment of the ethical standards and deontological criteria affecting this research (ref. 2710/2014).

52 Subjects and methods

2.1. Study population

The study population included up to 1,050 men and women of Caucasian ancestry who voluntarily attended to the Nutrigenetic Service located in community pharmacies in seven regions of Spain (Barcelona, Zaragoza, La Coruña, Pontevedra, Madrid, Granada and Málaga). The number of subjects involved in the present study are detailed in each corresponding chapter.

2.2. Intervention

The Nutrigenetic Service NS includes a program for weight loss. In the present work, the screened group in the intervention study (Chapter 5) included subjects who completed the first follow-up visit between 3 and 6 weeks after the baseline visit. Individual energy requirements were evaluated by computing resting energy expenditure according to the Harris-Benedict formula, and total energy expenditure according to the physical activity level as described elsewhere (395). Diets were designed to provide about 600 kcal/day less than the individually estimated total energy expenditure (restriction of 603 ± 15 kcal). The target macronutrient composition of the diets was: 20% of total energy from proteins, 30% from lipids and 50% from carbohydrates. In order to provide personalized nutrition based on both, the phenotype and the genotype, nutritionists instructed the participants individually about the type of foods they could eat based on a food exchange list system. For example, a subject with hypertension and genetic predisposition to high LDL cholesterol should follow a diet low in sodium, cholesterol and SFA.

2.3. Data collection

2.3.1. Anthropometric and body composition measurements

Anthropometric measurements were collected by trained nutritionists using a standardized protocol. Individuals were weighed with a digital scale (TANITA BF522W), wearing light clothes and no shoes. Height was measured using a portable stadiometer (Leicester TANITA) with subjects in barefoot. BMI was calculated dividing weight (kg) by the square of height (m). Body fat mass was determined by bioelectrical impedance using the TANITA BF522W model. Body fat distribution was evaluated by the measurement of waist and hip circumferences using a flexible and inextensible tape measure. Waist circumference was measured at the midway between the lower margin of the least rib and the top of iliac crest or according to the circumference at the level of the umbilicus if it was not possible to identify the least rib or the

53 Subjects and methods iliac crest; and hip circumference as the widest circumference over the greater buttocks. Waist-to-hip ratio and waist-to-height ratio were then calculated.

2.3.2. Blood pressure

Blood pressure was measured using the following standardized protocol with a validated automatic device (MIT Elite Plus, OMRON Healthcare, Hoofddorp, the Netherlands) and appropriately sized cuff (396). Measurements were carried out in the non-dominant arm, with the elbow at the level of the right atrium and with the subject in a sitting position. SBP and DBP were taken two times, separate of at least 10 min. The last measurement was used in the analysis, discarding the first one. If in the second reading the SBP or DBP were ≥140 or ≥90 mmHg, respectively, was performed a third measurement.

2.3.3. Dietary assessment

At baseline, diet information was collected by a validated FGFQ based on 19 food groups (397). Each food group comprised several foods based on a food exchange list system. Total energy intake and macronutrient composition were calculated using validated Spanish food composition tables (391–393).

2.3.4. Physical activity assessment

Physical activity was determined using a short 24-h physical activity questionnaire (395). Subjects were asked about the number of hours resting and practicing activities at work or at leisure time during a week day and a weekend day. Activities were divided in four groups according to intensity of effort: sedentary, low active, active and very active. Individual daily physical activity level was calculated multiplying the average time spent on each group of activities during the week and the weekend and the multiples of physical activity levels (395).

2.4. Genotyping

Samples from oral epithelial cells were collected (ORAcollect DNA®, DNAGenotek) in order to extract genomic DNA by QIAcube using QiAmp DNA Mini QIAcube Kit (Qiagen, Hilden, Germany), following the manufacturer procedures. The polymerase chain reactions (PCR) were carried out using the GeneAmp® PCR System 9700 thermal cycler (Applied Biosystems, Foster City, California) according to standardized laboratory protocols. PCR products were

54 Subjects and methods hybridized onto oligonucleotide probes attached to microspheres and labeled with streptavidin-conjugated phycoerythrin (MagPlex-TAG Microspheres). These beads were analyzed using Luminex® 100/200TM System, which is based on the principles of xMAP® Technology. Briefly, this method uncompressed polystyrene microspheres internally dyed with various ratios of spectrally distinct fluorophores, which are detected by a flow cytometry- based instrument (398). The genotyping was carried out by Progenika Biopharma (Bizkaia, Spain).

3. OBEKIT STUDY

The OBEKIT study (development of a nutrigenetic test for the prescription of body weight loss personalized diets) is a 10-months randomized, longitudinal and controlled intervention trial to evaluate the response to a two hypocaloric diets with different macronutrient composition based on the genetic background. The study lasted a total of 10 months divided in two sequential periods: one intervention period of 4 months where subjects were prescribed a hypocaloric diet and a maintenance period of 6 months in which subjects were prescribed a normocaloric diet in order to maintain the body weight loss during the first period.

The OBEKIT study protocol was approved by the Research Ethics Committee of the University of Navarra (ref. 132/2015) and was registered at Clinical Trials (clinical trial reg. no. NCT02737267). The study was performed in accordance to the ethical guidelines of the Declaration of Helsinki (389). All participants provided written informed consent after they received an information sheet and additional verbal explanation of the protocol.

3.1. Study population

The study population was recruited from October 2015 through February 2017 in the Metabolic Unit of the Centre for Nutrition Research of the University of Navarra. The recruitment of the participants was through the database from previous studies of the Metabolic Unit and local advertisements.

A group size of 200 volunteers was estimated to be necessesary in order to obtain a significant (p<0.05, 95% confidence interval) difference in weight loss between groups of 2 kg with a statistical power of 80% (b = 0.20). Given an estimated dropout rate of 30%, the sample size was fixed to 260 volunteers. In the present work, a total of 147 subjects were included.

55 Subjects and methods

Volunteers were randomly assigned to one of the two diets by a specific logarithm design for the study by MATLAB using stratified block randomization according to gender, age groups (18-40 and 41-70 years), ethnicity (Caucasian and Hispanic) and BMI (overweight, BMI 25-29.9 kg/m2; and obesity, BMI 30-40 kg/m2) (Figure 8). The inclusion and exclusion criteria are detailed (Table 23).

Assessed for eligibility n=174 Excluded n=27 Ineligible n=24 Declined to participate n=3 Randomized n=147

Low fat diet Moderately high protein diet 22% Fat / 18% Protein / 60% CH 30% Fat / 30% Protein / 40% CH n=72 n=75 Allocation up - n=53 n=54 16 weeks Follow

Figure 8. Flow-chart of participants in the OBEKIT study. Abbreviations: CH, Carbohydrates

Table 23. Inclusion and exclusion criteria of the OBEKIT sudy Inclusion criteria Adults: 18-70 y.o. BMI: ³25 kg/m2 and <40 kg/m2 Physical examination and vital signs normal, or is considered abnormal, but clinically insignificant by researcher In the case of individuals with chronic stable dose drug treatment and during the last 3 previous months at baseline, the investigator will assess their possible inclusion Exclusion criteria BMI <25 or ³40 kg/m2 Pregnant women Breastfeeding period. If artificial feeding until 6 months after birth Severe kidney and digestive system diseases Electrolyte disorders (disorders of sodium, potassium, calcium, chlorine, phosphorus, magnesium) Acute cardiovascular diseases Cancer Eating disorders Recent prescription drug treatment (without stable doses scheduled) Weight loss medications or others drugs that affect body weight (anti-psychotic or anti-depressant drugs, corticosteroids) Some type of cognitive / psychic impairment Subjects in which poor collaboration or, in the investigator's opinion, have difficulty following the procedures of the study is foreseen Lack of commitment (at the discretion of the investigator) with the intervention, suspected non-compliance, or real difficulties to follow the development of the study. Abbreviations: BMI, Body mass index

56 Subjects and methods

3.2. Intervention

Energy requirements were evaluated individually from resting energy expenditure, according to the Mifflin formula, multiplied for physical activity level calculated by a short 24-h physical activity questionnaire (395,399,400). Diets presented the following target macronutrients: low-fat diet: 60% of total energy from carbohydrate, 18% from protein and 22% from fat; and moderately high-protein diet: 50% of total energy from carbohydrate, 30% from protein and 30% from fat. Prescribed diets provided a 30% restriction of the total energy expenditure estimated for each subject. No initial diets had less than 1200 kcal/day.

Trained nutritionists designed both diets based on a food exchange system. Participants were given a menu template with the number of exchanges of each food group for each meal, the list of food exchanges, and structured daily meal plans for 2 weeks. The exchange system offered to the participants the most flexibility in diet planning once the concept was learned. In addition, volunteers were instructed to weight all the food they consumed and all of them received instructions orally and in writing.

Participants were encouraged to maintain their usual physical activity, which was controlled by a short 24-h physical activity questionnaire administered at the beginning and at the end of the study (393). In addition, volunteers were asked to collect the information of a pedometer at the beginning, at the follow-up visit and at the end of the study.

3.3. Data collection

3.3.1. Anthropometric and body composition measurements

Anthropometric and body composition measurements were taken at the beginning and at the end of the study in fasting conditions with the subjects in their underwear. Body weight was assessed using the TANITA BC-418 (Tanita, Tokyo, Japan) and height was measured using a wall-mounted stadiometer. BMI was calculated dividing weight (kg) by the square of height (m). Waist circumference was measured at the midway between the lower margin of the least rib and the top of iliac crest or according to the circumference at the level of the umbilicus if it was not possible to identify the least rib or the iliac crest. Hip circumference was assessed at the widest circumference over the greater buttocks. Waist and hip circumferences were taken by a stretchable tape measure. Body composition was analyzed by bioelectrical impedance using the TANITA BC-418 (Tanita, Tokyo, Japan), and by DXA scan (DXA Lunar Prodigy, GE Medical Systems, Madison, WI, USA) by trained and certified observers.

57 Subjects and methods

3.3.2. Blood pressure

Blood pressure was measured using a validated automatic device (MIT Elite Plus, OMRON Healthcare, Hoofddorp, the Netherlands) as described by the manufacturer (396). At each visit, 2 readings were obtained with the elbow at the level of the right atrium and in the seated position, separate of at least 10 minutes. The average of the 2 measurements was used in the analysis.

3.3.3. Dietary assessment

Dietary habits at baseline were assessed by a validated semi-quantitative FFQ of 137 food items and by a 3-day weighed food record (two weekdays and one weekend day) (401–403). Compliance to the recommended diet of the participants was conducted taking into account 3- day weighed food record at 8th week and at the end of the intervention period. Total energy intake and nutrient content were determined using Spanish food composition tables (391,392).

3.3.4. Physical activity assessment

Physical activity was evaluated applying a combination of methods: a validated physical activity questionnaire, a short 24-h physical activity questionnaire, previously defined in the present document, and a pedometer (HJ-321-E, OMRON Healthcare, Hoofddorp, the Netherlands) (395,404).

3.3.5. Blood samples

Blood samples were collected at baseline and at the end of the intervention as well as at the maintenance period after a 12-h overnight fast. Fasting serum plasma glucose and lipid concentrations were measured in a Pentra C-200 auto analyzer (HORIBA ABX, Madrid, Spain) with specific kits. LDL cholesterol was calculating according to the Friedewald equation (405). Meanwhile fasting plasma insulin concentrations were measured with a double antibody radio-immunoassay (Insulin RIA 100, Kabi-Pharmacia, Uppsala, Sweden).

58 Subjects and methods

3.4. Genotyping

For genotyping epithelial buccal cells were collected using a liquid based kit (ORAcollect-DNA, OCR-100, DNA genotek, Ottawa, Canada). Genomic DNA was extracted with the Maxwell 16 Buccal Swab LEV DNA Purification Kit in the Maxwell 16 instrument (Promega, Madison, USA). Genotyping of ADCY3 rs10182181 was performed by Next-Generation Sequencing using a pre- designed SNP panel (Ion AmpliSeq Custom NGS DNA Panels, Thermo Fisher Scientific Inc, Waltham, MA, USA), which was validated in the Ion Torrent PGM (Thermo Fisher Scientific Inc, Waltham, MA, USA). Data were analyzed with the Torrent Variant Caller plugin for the Ion Torrent Sequencing platform and R software.

4. POUNDS LOST TRIAL

The POUNDS Lost, is a 2-years randomized clinical trial designed to compare the effects on body weight of energy restricted diets that differed in their targets for intake of macronutrients and otherwise followed recommendations for cardiovascular health (243). Although the primary outcome of the study was the change in body weight over a period of 2 years, weight loss typically is greatest after 6 months of the initiation of the diet. Because of that in the present work we studied changes in metabolic traits after 6 months and 2 years of the dietary intervention.

The study was approved by the human subjects committee at each participating institution and by a data and safety monitoring board appointed by the National Heart, Lung and Blood Institute. The study was registered at Clinical Trials (clinical trial reg. no. NCT00072995). All participants provided written informed consent.

4.1. Study population

The POUNDS Lost trial was conducted from October 2004 through December 2007 at two sites: Harvard School of Public Health and Brigham and Women’s Hospital in Boston, MA, and the Pennington Biomedical Research Center of Louisiana State University System, Baton Rouge, LA. The data coordinating center was at Brigham and Women’s Hospital. The primary means of recruitment were mass mailings (the primary sources of mailing lists were commercial vendor and local governments, for list of registered voters or drivers). Secondary method included advertisements on buses and subways, worksite and newspaper, distribution of recruitment flyers and mailings to local healthcare center and businesses.

59 Subjects and methods

The study population included 811 overweight or obese (BMI: 25-40 kg/m2) individuals who were randomly assigned to one of four diets by the data manager at the coordinating center, upon request of a study dietitian, after confirming, by computer program, that all screening activities had occurred and that the participant met all eligibility criteria (Figure 9). Diet group assignments were stratified by site with varying block sizes to ensure a balance at each site. The inclusion and exclusion criteria are detailed in Table 24.

Assessed for eligibility n=1638 Excluded n=827 Ineligible n=461 Declined to participate n=366 Enrollment Randomized n=811

Low fat, average protein diet Low fat, high protein diet High fat, average protein diet Low fat, high protein diet 20% Fat / 15% Protein / 65% CH 20% Fat / 25% Protein / 55% CH 40% Fat / 15% Protein / 45% CH 40% Fat / 25% Protein / 35% CH n=204 n=202 n=204 n=201 Allocation up - n=169 n=157 n=151 n=168 2 years Follow

Figure 9. Flow-chart of participants in the POUNDS Lost trial. Abbreviations: CH, carbohydrates

Table 24. Inclusion and exclusion criteria of the POUNDS Lost trial Inclusion criteria Adults: 30-70 y.o. BMI: ³25 kg/m2 and <40 kg/m2 Type 2 diabetes controlled with diet Hypertension or hyperlipidemia treated with diet or drugs Exclusion criteria Diabetes treated with oral medications or insulin Serious gastrointestinal disease Alcohol or drug abuse Eating disorders Unstable or recent onset of cardiovascular disease Weight loss medications or others drugs that affect body weight (anti-psychotic or anti-depressant drugs, corticosteroids) Hypothyroidism defined by abnormal thyroid stimulating hormone Urinary microalbumin >100 µg/g creatinine Unstable dose of medication of hyperlipidemia, hypertension or psychiatric disorder Insufficient motivation as assessed by interview and questionnaire Other serious illness Abbreviations: BMI, Body mass index

60 Subjects and methods

4.2. Intervention

Estimated energy requirements were calculated from resting energy expenditure, which was estimated by Delta Trac II metabolic cart, plus activity coefficient. Each participant’s diet prescription represented a caloric restriction of 750 kcal/day. The target percentages of energy derived from fat, protein and carbohydrate in the four diets were: 20, 15 and 60%; 20, 25 and 55%; 40, 15 and 45%; and 40, 25 and 35%, respectively. Thus, the 4 diets constituted a 2-by-2 factorial design: two diets were low-fat (20%), two diets were high-fat (40%), two diets were average-protein (15%) and two diets were moderately-high-protein (25%). The goals for kinds of fatty acids were: SFA 8% for each group; MUFA 6% for low-fat diet and 22% for high-fat diet; PUFA 6% for low-fat diet and 10% for high-fat diet. The goal for dietary fiber and for dietary cholesterol were 20 g/day minimum and 150 mg/1000 kcal, respectively for all groups. Study participants were encouraged to use low glycemic index carbohydrates-rich foods and a multivitamin with calcium 200-250 mg/day.

The 4 diets were designed using the ADA exchange system. Structured daily meal plans 2-week blocks were given to the participants. Moreover, food shopping lists and easy-to-prepare recipes were provided.

Physical activity goals were established for sedentary participants, gradually increasing from 30 minutes/week of moderate intensity exercise to 90 minutes/week during the first 6 months. This goal remained constant over the next 18 months of the trial. Meanwhile, participants who were accustomed to or desired to achieve a higher exercise goal were encouraged to do so. Time of exercise was monitored using self-monitoring forms.

4.3. Data collection

4.3.1. Anthropometric and body composition measurements

Body weight and waist circumference were measured the morning before breakfast on two nonconsecutive days at baseline, 6 and 24 months; and on a single day at 12 and 18 months. Height was measured on two nonconsecutive days at the baseline examination. Individuals were weighed with calibrated hospital scales. Waist circumference was measured 4 cm above the iliac crest using a non-stretchable tape measure. BMI was calculated dividing weight (kg) by the square of height (m) (243).

Body composition was analyzed by a DXA scan using a Hologic QDR 4500A (Hologic, Inc.) after an overnight fast in a random sample of 50% of the total study population by trained and

61 Subjects and methods certified observers (406). Total fat mass (kg), total lean mass (kg), percentage of whole body fat mass and percentage of trunk fat were obtained at baseline, 6 months and 2 years of the intervention.

4.3.2. Blood pressure

Blood pressure was measured on two days at screening visit and at 6, 12 and 24 months using an automatic device (IntelliSense Professional Digital Blood Pressure Monitor, HEM907XL, OMRON Healthcare). At each visit, 3 readings were obtained in the seated position by trained and certified observers. The average of the 3 measurements was used in the analysis.

4.3.3. Dietary assessment

To evaluate the adherence to the dietary intervention program, dietary intake was assessed in a random sample of 50% of the participants by a review of the 5 days diet record at baseline and by 24-h recall during a telephone interview on 3 nonconsecutive days at 6 months and 2 years of follow-up.

4.3.4. Physical activity assessment

Habitual physical activity was estimated using the Baecke physical activity questionnaire a valid and reliable 16-item self-report inventory (407). The questionnaire was administered at baseline, 12 months and 2 years of the intervention. Self-reported physical activity was also tracked through the computer tracking system during each week of the study.

4.3.5. Blood samples

Blood samples were collected in the fasting state at baseline and at 6 and 24 months of follow- up. Serum was aliquoted and stored at -80ºC at each clinical site until the analyses. Bioquemical analyses were performed at the Clinical Laboratory at Pennington. Levels of serum lipids (triglycerides, total cholesterol and HDL cholesterol) were analyzed using the Synchron CX7 (Beckman Coulter, Brea, CA). LDL cholesterol was calculated for each participant according to the Friedewald equation (405). However, when triglyceride concentration exceeded 400 mg/dL, LDL cholesterol was measured directly by Synchron CX7 (Beckman Coulter, Brea, CA).

62 Subjects and methods

4.4. Genotyping

DNA was isolated from the buffy coat fraction of centrifuged blood using the QIAmp Blood Kit (Qiagen, Chatsworth, CA). For the present study, the rs10830963 MTNR1B genetic variant was genotyped using the OpenArrayTM SNP Genotyping System (BioTrove, Woburn, MA). Replicated quality control samples (10%) were included and genotyped with >99% concordance.

5. NUGENOB STUDY

The NUGENOB study is a randomized, parallel, two-arm, open-label 10-week dietary intervention of two hypo-energetic diets (high- versus low-fat diet) to examine if there is an interaction between the nutrient composition of the diet, specifically the fat content, and obesity related genes in response to the dietary treatment (246).

The study protocol was approved by the Ethics Committee of each centre/country and was registered at ISRCTN registry (ISRCTN25867281). All subjects gave written informed consent after volunteers were informed about the nature and risks of the experimental procedure.

5.1. Study population

The recruitment of the subjects was undertaken from May 2001 until September 2002 in eight sites in seven European countries: United Kingdom (England), The Netherlands, France (two centers), Spain, Czech Republic, Sweden and Denmark.

In the NUGENOB study it was planned to recruit 100 subjects from each center of the seven centers and 50 from one center. This sample size would allow the study to detect 0.7 kg difference in weight loss, assuming a standar deviation of 4 kg, a significance p-value of 0.05 and a statistical power of 0.90. At the end the study population encompassed a total of 771 individuals who were randomly assigned to one of the two diets by contacting the coordinating center at each allocation (Figure 10). Stratified block randomization was used with center, gender and three groups of age (20-29, 30-39 and 40-50 years old) as strata. The randomization list was computer generated. The inclusion and exclusion criteria have been detailed (Table 25).

63 Subjects and methods

Randomized n=711

Low fat diet High fat diet 20-25% Fat / 15% Protein / 60-65% CH 40-45% Fat / 15% Protein / 40-45% CH n=389 n=382 Allocation up -

weeks n=336 n=312 10 Follow

Figure 10. Flow-chart of participants in the NUGENOB study. Abbreviations: CH, carbohydrates

Table 25. Inclusion and exclusion criteria of the NUGENOB study Inclusion criteria Adults: 20-50 y.o. BMI: ≥30 kg/m2 White European (by self-report) Pre-menopausal (women) Exclusion criteria Weight change >3 kg within the 3 months before the beginning of the study Drug-treated hypertension, diabetes or hyperlipidemia Untreated thyroid disease Use of anorexigenic agents Anti-epilectic or anti-parkinsonian drugs Surgically treated obesity Pregnancy Use of barbituates, benzodiasapines, beta-blockers, butyrophenones, carbonic anhydrase inhibitors, diuretics, dopamine reuptake inhibitors, digoxin, fibrates, fish oil supplement, glucocorticoids, immunosuppressives, insulin, laxatives, monoamine oxidase inhibitors, niacin (>150 mg per day), nicotine, oral hypoglycemics, orlistat, phenothiazines, selective serotonin reuptake inhibitors, serotonin and norepinephrine reuptake inhibitors, statins, thyroid hormone, triamterene, tricyclic antidepressants, warfarin or zonisamide. Alcohol or drug abuse Participation in other simultaneous ongoing trials Abbreviations: BMI, Body mass index

5.2. Intervention

The target macronutrient composition of the two diets were: low-fat diet 20-25% of total energy from fat, 15% from protein, and 60-65% from carbohydrate; high-fat diet 40-45% of total energy from fat, 15% from protein and 40-45% from carbohydrate. Both diets were designed to provide 600 kcal/day less than the individually estimated energy requirement based on an initial resting metabolic rate multiplied by 1.3. Subjects were given oral and written instructions to minimize differences between the two diets in type of fat, amount and type of fiber, type of carbohydrate, fruit and vegetables, and meal frequency. Subjects were requested to abstain from alcohol consumption.

64 Subjects and methods

The diets were designed using and exchange system, and instructions were provided on either a template or an exchange system. Dietary advice reflected local customs of each country. The dietary instructions were reinforced weekly with the dietitian.

5.3. Data collection

5.3.1. Anthropometric and body composition measurements

Body weight and height were measured with calibrated scales and calibrated stadiometers, respectively. Waist and hip circumferences were measured with the participant wearing only non-restrictive underwear. The mean of the three measurements was recoded for each variable. BMI was calculated as weight in kilograms divided by height in meters squared. Fat mass and fat-free mas were estimated by bioelectrical impedance (Bodystat®; QuadScan 400, Isle of Man, British Isles).

5.3.2. Dietary assessment

To assess the dietary adherence across the intervention, a 3-day weighed food record of two weekdays and one weekend day was performed at baseline and at last week of the intervention. Moreover, one-day weighed food records were completed in the 2nd, 5th and 7th weeks. The dietary records were analyzed using the food nutrient database routinely used in each center.

5.3.3. Blood samples

Venous samples were collected after an overnight fast of 12-h and after participants rested supine for 15 min. Fasting plasma glucose and lipid concentrations were measured using standard enzymatic techniques on a COBAS FARA centrifugal spectrophotometer (Roche Diagnostica, Basel Switzerland; glucose HK 125, ANX Diagnostics, Montpellier, France; triglycerides, Sigma, St Louis, MO, USA; total cholesterol, cholesterol 100, ABX Diagnostics Montpellier, France; HDL cholesterol, HDL-C, Roche, IN, USA). Fasting plasma LDL cholesterol was calculating according to the Friedewald equation (405). Fasting plasma insulin concentrations were measured with a double antibody radio-immunoassay (Insulin RIA 100, Kabi-Pharmacia, Uppsala, Sweden). IR was estimated by homeostasis model assessment of IR (HOMA-IR) using the following equation: [fasting insulin (µL/mL)] x [fasting glucose (mg/dL) / 18.01] / 22.5 (408–410). b cell function was estimated by homeostasis model assessment of ß

65 Subjects and methods cell function (HOMA-B) as follows: [20 x fasting insulin (µL/mL)] / {[fasting glucose (mg/dL) / 19.01] – 3.5} (408–410).

5.4. Genotyping

DNA was extracted from buffy coats in the Steno Diabetes Center (Copenhage). Extracted DNA samples were diluted in Tris/EDTA buffer to a stock DNA solution of 100 ng µL-1 and a working DNA solution of 10 ng µL-1. Stock solutions were preserved at -80ºC meanwhile working solutions were stored at 4ºC. DNA samples were kept and handled in locations free of contaminating polymerase chain reaction products. For the current study the rs1440581 PPM1K genetic variant was successfully genotyped in 757 individuals (99.3% genotyping success rate) at LCG group by the KASPTM genotyping assay (United Kingdom).

6. STATISTICAL ANALYSES

Chi-squared test was performed for comparison of categorical variables (e.g. gender, ethnicity) across genotypes or GRS groups. Moreover, deviation from Hardy-Weinberg equilibrium (HWE) was tested by chi-squared test (411). General linear models for continuous variables (analysis of the variance -ANOVA- or analysis of the covariance -ANCOVA- when was adjusted for cofounder variables) were performed for comparison of continuous variables across gender or genotypes (e.g. age, energy intake). Cualitative variables were expressed as n (%) and quantitative variables as mean (standard deviation (SD)).

The validity of the FGFQ was analyzed using different methods. To determine the strength of the association between both methods the Pearson’s or Spearman’s correlation coefficients were applied depending on the normality of the variables. The degree of agreement between the two questionnaires was assessed by the intraclass correlation coefficient (ICC) and the Bland-Altman method known as limits of agreement (LOA) (412,413). Finally, to know the ability of the FGFQ to categorize the volunteers in quartiles of energy intake or macronutrients when compared with the reference method the cross classification analysis was used (414).

Multiple linear regression analyses were used to evaluate the association between a specific polymorphism or GRS and the outcome as continuous variable of each chapter (e.g. body fat, weight loss, SBP) after adjusted for covariates. Multiple logistic regression analyses were performed when the outcome was dichotomous (e.g. obesity, hypertension) once adjusted for

66 Subjects and methods confounder variables. In both analyses, the interaction term was included in the model to evaluate gene-environment interactions (e.g. GRS x high-/low-carbohydrate intake).

The GRS was calculated assuming that each SNP acts independently and contributes equally to the risk of obesity in an additive manner, as has been previously reported (415,416). Genotypes were coded as 0, 1 or 2 according to the number or risk alleles for each variant. The GRS was computed by summing the risk alleles across the 16 selected SNPs for each individual. The discriminative power of the GRS was tested by the area under the receiver operating characteristics curves (ROC AUCs) (417).

Least angle regression (LARS) analysis was selected as model selection technique (418). Because LARS algorithm is designed for linear regression with continuous or binary covariates, polymorphisms were re-coded in binary variables according to the association between each polymorphism and BMI tested by using dummy linear regression models. In those cases where there was no significant association and due to the limited frequency of the variant allele, homozygotes of the minor allele (aa) and heterozygotes (Aa) were grouped. Stagewise regression and Lasso were also performed to verify the selection of the independent variables established by LARS (418). Bootstrapping was performed to internally validate the regression model by constructing a number of resamples (k Z 100) of the dataset.

All statistical analyses were conducted using STATA/SE version 12.0 (StataCorp, College Station, TX, USA). Tests were considered statistically significant at p value <0.05. These analyses are explained with more detailed in each corresponding chapter.

67

RESULTS

INTRODUCTORY RESEARCH

CHAPTER 1

Single-nucleotide polymorphisms and DNA methylation markers associated with central obesity and regulation of body weight

Goni L.1,2, Milagro F.I. 1,2,3, Cuervo M.1,2,3, Martínez J.A.1,2,3

1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Navarra, Spain 2 Centre for Nutrition Research, University of Navarra, Navarra, Spain 3 Biomedical Research Centre Network in Physiopathology of Obesity and Nutrition (CIBERobn), Institute of Health Carlos III, Madrid, Spain

Nutr Rev. 2014 DOI 10.111/nure.12143 Impact factor (2014): 6.076 4/77 Nutrition & Dietetics, Q1

Results

75 Chapter 1

ABSTRACT

Visceral fat is strongly associated with the development of specific obesity-related metabolic alterations. Genetic and epigenetic mechanisms seem to be involved in the development of obesity and visceral adiposity. The aims of this review are to identify the single-nucleotide polymorphisms related to central obesity and to summarize the main findings on DNA methylation and obesity. A search of the MEDLINE database was conducted to identify genome-wide association studies, meta-analyses of genome-wide association studies, and gene-diet interaction studies related to central obesity, and, in addition, studies that analyzed DNA methylation in relation to body weight regulation. A total of 8 genome-wide association studies and 9 meta-analyses of genome-wide association studies reported numerous single- nucleotide polymorphisms to be associated with central obesity. Ten studies analyzed gene- diet interactions and central obesity, while 2 epigenome-wide association studies analyzed DNA methylation patterns and obesity. Nine studies investigated the relationship between DNA methylation and weight loss, excess body weight, or adiposity outcomes. Given the development of new sequencing and omics technologies, significantly more knowledge on genomics and epigenomics of obesity and body fat distribution will emerge in the near future.

Keywords Visceral fat, GWAS, gene-diet interaction, EWAS, DNA methylation

76 Results

1. INTRODUCTION

Obesity can be defined as a complex disease characterized by an abnormal or excessive accumulation of body fat resulting from an imbalance between the energy consumed and the energy expended.1 Other factors, however, could be involved, such as lack of physical activity, endocrine disorders, social determinants, or genetics.2–5 Moreover, obesity is related to an increased incidence of a number of physiopathological conditions, including type 2 diabetes, cardiovascular disease, dyslipidemia, hypertension, sleep apnea, polycystic ovarian syndrome, osteoarthritis, and some forms of cancer, among others.6 In this context, it is generally accepted that visceral fat, in contrast to total body fat, is associated specifically with the development of obesity-related metabolic disorders.7 Central obesity can be indirectly assessed by anthropometrical measurements, such as waist circumference and waist-hip ratio, or bymore complex methods, such as bioelectrical impedance analysis (ViScan; Tanita Corporation, Arlington Heights, IL), dual energy X-ray absorptiometry, and computed tomography (CT).8,9 About 40–70% of the interindividual variation in polygenic obesity,which results from the combined eff ect of genetics and environmental factors, can be attributed to genetic causes.5 The analysis of genetic diff erences is based on the study of single-nucleotide polymorphisms (SNPs). The occurrence of polymorphisms can have functional consequences such as changes in the amino acid sequences, in the mRNA transcript stability, and in the transcription factor binding affi nity.10 The first investigations to assess genes that predispose to obesity appeared in the mid-1990s and were based on candidate gene studies.11 Later, following the sequencing of the human genome by the Human Genome Project and the determination of the common patterns of DNA sequence variation by the Haplotype Map Project, genome-wide linkage studies were used to search for genes that predispose to obesity.12,13 The progress of high-throughput genotyping methods such as genome-wide association studies (GWAS) has allowed the identification of many new loci associated with polygenic diseases.14 These techniques have already revealed genomic variants that predispose to several disorders, thereby providing new targets for prevention, diagnosis, or therapy.15 In fact, genetic test panels could be useful for developing personalized nutrition guidance based on genetic predisposition.16 Most of the SNPs analyzed by companies or laboratories are located in genes related to general obesity, body weight regulation, diabetes, hyperglycemia, hyperlipidemia, and thrombosis, but research on visceral fat deposition is scarce, despite the association between visceral fat and metabolic diseases.16 Although specific genomic variations can explain a percentage of the risk of human obesity, epigenetic, heritable changes in gene expression without changes in the DNA sequence also may be

77 Chapter 1 involved.17 The main epigenetic mechanisms are DNA methylation and histone modifications, including methylation, acetylation, ubiquitination and sumoylation of lysine, phosphorylation of serine and threonine, and methylation of arginine.18 Technological advances in molecular biology and genetics have facilitated the study of the epigenome. Thus, histone modifications at the whole genome scale are currently examined using chromatin immunoprecipitation followed by genomic tiling microarray hybridization methodology (ChIP/chip).19 DNA methylation can be analyzed by genome-wide microarrays after bisulfite conversion, by methylation-sensitive restriction digestion assays, or by immunoprecipitation-based enrichment assays.20 In addition to GWAS, the concept of epigenomewide association study (EWAS) has been developed. Sequencing and array-based profiling technologies have been used for this type of analysis.21 However, as DNA methylation patterns at specific CpG sites are tissue specific and can vary over time within an individual, the results must be validated in multiple populations.22 The two aims of this review are to summarize the GWAS and the meta- analyses of published GWAS related to central obesity or visceral fat and to outline the main findings on DNA methylation and its relationship to the obesity phenotype.

2. METHODS

This review is presented into two parts: 1) genomics and central obesity and 2) DNA methylation and obesity. An extensive search of the PubMed electronic database to identify human studies published in the English language until August 2013 was performed. Since there were two main topics of the review, the terms used to the search of publications differ, as shown below.

For GWAS focused on central obesity, including meta-analyses of GWAS, the following search terms were used: [GWAS OR meta-analysis GWAS] AND [central obesity OR visceral fat OR body fat OR obesity traits OR waist circumference OR waist-hip ratio]. Studies were excluded if there was no reference to central obesity (waist circumference, waist-hip ratio, visceral adipose tissue or ratio of visceral adipose tissue to subcutaneous adipose tissue). The 17 studies included were reviewed in depth to find SNPs associated with waist circumference, waist-hip ratio, visceral adipose tissue, and ratio of visceral adipose tissue to subcutaneous adipose tissue. Table 123-30 summarizes the SNPs reported by GWAS, and Table 231-39 shows those SNPs reported by meta-analyses of GWAS. There is some ambiguity about the most suitable threshold for claiming genome-wide significance.40-42 In fact, among the reviewed studies, some authors proposed their own level of significance. Thus, it was decided to include

78 Results in Table 1 only those variants reaching significance levels of P<10-6. The rest of the polymorphisms described in the literature are presented in Table S1 (available online in the Supporting Information). In contrast, Table 2 (meta-analyses) includes polymorphisms reaching significance levels of P<10-8 and in Table S2 (available online in the Supporting Information) shows those variants with significance levels between P<10-8 and P<10-6.

The terms used to search for publications about gene-diet interactions and central obesity were as follows: (gene OR genetics OR gene interaction) AND (diet OR fat intake OR saturated fatty acids [SFA] intake OR polyunsaturated fatty acids [PUFA] intake OR monounsaturated fatty acids [MUFA] intake) AND (central obesity OR visceral fat OR body fat OR WC OR WHR). Additionally, other published reports were obtained by cross-matching references of selected articles. Again, studies were excluded if no outcome was related to central obesity. Table 343-52 gives an overview of gene-diet interaction studies associated with central obesity.

To search for studies about DNA methylation and obesity, the following terms were used: (EWAS OR Genome wide DNA methylation OR epigenetics OR methylation OR methylation array OR methylation microarray) AND (obesity OR body fat OR adiposity OR weight loss OR weight gain).

All the studies were screened, and the heading “DNA methylation and obesity” was analyzed under three subheadings: “EWAS in obesity”, “DNA methylation markers and body weight control” and “Methylation at early age and development of excess weight or adiposity”. Table 453,54 summarizes the 2 EWAS on obesity. Studies about methylation patterns after a weight- loss intervention are included in Table 5.55-60 Some of the studies used a methylation microarray; in these cases, Table 5 shows the CpG sites validated by a second method in these cases. Studies on methylation patterns at early age are discussed only in the text.

3. GENOMICS AND CENTRAL OBESITY

3.1. Genome-wide association studies

In recent years, a large number of GWAS on obesity have been performed in different populations.61 The current review is focused only on GWAS related to central obesity (Table 1, Table S1).

The first GWAS based on central-obesity-related traits was published by Chambers et al.26 in 2008. The authors carried out a 2- stage GWAS with a combined analysis (stage 1 and stage 2), in which the most robust statistical significance was identified for the association between

79 Chapter 1

SNPs located close to the gene MC4R and waist circumference, specifically the SNP rs12970134. In this context, MC4R, which is a hypothalamic regulator of energy balance, was found to be related to obesity. More than 100 variants in human MC4R have been reported up to date.62 A later study in Chinese children, confirmed the association of a specific SNP located in MC4R (rs17782313) with central obesity risk.63

The article by Cho et al.25 in 2009 reported new SNPs associated with waist-hip ratio. The genome-wide results obtained from an East Asian population revealed 8 new SNPs. Two of these SNPs were genotyped in a second population of the same origin. The result of the 2 combined analyses confirmed the relationship between rs2074356 and waist-hip ratio. This SNP is located in the 24th intron of the C12orf51 transcript, whose function has not yet been elucidated.25

Polasek et al.29 selected a sample from an isolated population, Korčula, Croatia, in order to decrease genetic and environmental heterogeneity among individuals. With regard to central body fat distribution, they measured waist circumference and showed that it was related to 2 SNPs in different genes (MAX and SEZ6L2) in this GWAS. The role and function of these 2 genes in human body fat regulation is not known.29

Norris et al.28 reported the first GWAS in combination with a technological method to measure visceral adipose tissue and subcutaneous adipose tissue. Axial images at the L4-L5 disc spaces were obtained by CT. They performed a multistage design of GWAS. SNPs with evidence of association in stage 1 were replicated in a larger sample, and the most significant association was rs7086207, located in a non-genic region (165.8 kb from AK123440 and 35.5 kb from MYO3A). The follow-up stage revealed a new possible candidate gene for central adiposity, RGS6. The RGS6 gene encodes a member of the RGS family of proteins, which is involved in the modulation of opioid receptors.64 Variants near or within this gene may affect the modulation of opioids, cortisol secretion and eating behavior.28

A case-control GWAS was performed by Wang et al.24 in 2011, obese adults (body mass index [BMI]≥35 kg/m2) were compared with normal weight adults (BMI≤25 kg/m2). The analysis demonstrated that a specific SNP in the NRXN3 gene was strongly associated with waist-hip ratio. Furthermore, NRXN3, expressed in the brain, has been associated with addictions such as alcohol and smoking, and eating disorders have been ascribed to the hypothalamic part of the brain.65-67 Therefore, polymorphisms near or within the gene NRXN3 may affect the brain function and eating behavior. This gene has also been associated with waist circumference by a meta-analysis of GWAS.39

80 Results

In the same year (2011), Croteau-Chonka et al.27 performed a GWAS in Filipino women. The strongest evidence of an association with waist circumference was observed with rs1440072 in the KCNE4 gene, which was also associated with BMI. The KCNE4 gene encodes a member of the potassium channel, which inhibits the KCNQ1 and other voltage-gated potassium channels.68 KCNQ1, which is expressed in adipose tissue, has been associated with type 2 diabetes and might have an effect on lipid metabolism.69

In 2012, Kristiansson et al.23 analyzed loci associated with manifestations of the metabolic syndrome. The authors observed an association between waist circumference and the rs9940128 SNP of the FTO gene. This gene, which encodes an enzyme that has been identified as a mRNA demethylase, is functionally connected with the regulation of IRX3 expression and is the most relevant polygene for obesity.70 A wide variety of FTO polymorphisms have shown a strong and highly significant association with obesity-related traits and also with metabolic disturbances related to excess body weight, such as cardiovascular disease and type 2 diabetes.63,71-75

Plourde et al.30 studied the relationship between several SNPs located in the LRRFIP1 gene and 2 central-obesity-related traits: waist circumference and visceral adipose tissue. The measurement of visceral adipose tissue was performed by CT, according to Norris et al.28 The SNP most strongly associated with visceral adipose tissue was rs11680012. They hypothesized that LRRFIP1 could be involved in toll-like receptors signaling pathway, which plays a significant role in inflammation.30,76

3.2. Meta-analysis of genome-wide association studies

In order to increase the statistical power of research, the results of different GWAS can be combined in a meta-analysis (Table 2, Table S2).

In 2009, Heard-Costa et al.39 conducted one of the first meta-analysis of GWAS. In the first stage, a meta-analysis of 8 GWAS reported that a total of 163 SNPs (in 22 genes) were associated with waist circumference. In the second stage, the authors selected 48 SNPs in order to conduct an in silico exchange study. Following this analysis, the results were meta- analyzed and showed that the SNP rs10146997 of the gene NRXN3 was associated with waist circumference, in accordance with the results of the GWAS of Wang et al.24 This gene has been previously implicated in studies of addiction and reward behavior.39

81 Chapter 1

A second meta-analysis was published by Lindgren et al.35 The initial meta-analysis (stage 1) examined a total of 2,573,738 SNPs in 16 GWAS to identify those SNPs most strongly associated with anthropometrical measurements of body fat distribution. A total of 26 SNPs were selected for 2 further different follow-up analyses: in-silico and de novo genotyping (stage 2). Finally, the joint analysis confirmed 2 genome-wide significant loci for waist circumference mapped near or within TFAP2B (rs987237) and MSRA (rs545854) genes and 1 loci for waist-hip ratio near LYPLAL1 (rs2605100). Later, each of these genes was found to be related to other anthropometric and metabolic traits.77 The TFAP2B gene, expressed mainly in adipose tissue, encodes for the transcriptional factor activating enhancer binding protein 2b (AP-2b) and has been previously associated with BMI and insulin resistance.71,78,79 Most of the variants previously implicated in obesity exert their effects on adiposity at the hypothalamic level. However, TFPA2B could be involved in a global adipocyte response to positive energy balance.35 The function of the MSRA gene is the repair of oxidative damage to proteins by the enzymatic reduction of methionine sulfoxide to methionine.80 The relationship between the MSRA locus and adiposity is unclear, but this gene has been reported to be involved in oxidative stress.81 Lindgrend et al.35 proposed TNKS, a gene located near MSRA, which interacts with insulin-responsive aminopeptidase in GLUT4 vesicles in adipocytes, as an alternative candidate gene for this association. The authors found that the SNP rs2605100, a locus near LYPLAL1, was associated with waist-hip ratio in women only. This gene encodes the lysophospholipase-like 1 protein, which acts as a triglyceride lipase and has been reported to be upregulated in visceral adipose tissue and subcutaneous adipose tissue of obese individuals.82

The study by Heid et al.32 began with a meta-analysis of 32 GWAS with the aim of identifying loci potentially associated with waist-hip ratio adjusted for BMI. This study evaluated 16 SNPs in 29 additional independent studies using 2 follow-up studies: in silico and de novo genotyping. The joint analysis of the 2 stages revealed genome-wide significant associations for 14 loci. The association between waist-hip ratio and the rs6905288 SNP of the VEGFA gene is of biological interest because this gene has been identified as a mediator of adipogenesis, involved in the angiogenesis process.83,84 The mechanism by which the rs10195252 SNP of the GRB14 gene is associated with visceral adipose tissue is still unknown. However, results of mRNA expression experiments indicated that this locus could be involved in the regulation of fat distribution.85 Consistent with this finding, GRB14 was subsequently found to be associated with triglycerides, insulin and high-density lipoprotein cholesterol (HDLc) levels.38 Available evidence supports the developmental genetic origin of patterns of fat accumulation and

82 Results distribution.86 TBX15 has been considered one of the developmental genes implicated specifically in visceral fat.32,85 The role of the other genes related to waist-hip ratio at genome- wide significance level on the development of body fat distribution has not yet been determined. Nevertheless, several replication studies confirmed this association as well as the putative relationship with metabolic-related traits.87,88

In 2011, 2 meta-analyses of GWAS were published. The study performed by Kraja et al.34 tested the hypothesis that genes with pleiotropic effects may be responsible for some pairs of metabolic traits implicated in metabolic syndrome. The meta-analysis of 7 GWAS identified a total of 12 SNPs associated with binary pairwise traits related to central obesity (waist circumference-triglycerides, waist circumference-HDLc and waist circumference-glucose) at a genome-wide significance level. The most significant association with a central adiposity measurement was the variant rs2266788 in APOA5 with waist cricumference-triglycerides. APOA5 encodes a protein found in chylomicrons, which regulates the plasma triglyceride levels, as well as very low-density lipoprotein and HDLc.89 The CETP gene, associated with the binary trait HDLc-waist circumference, has a functional role in the transfer of cholesterol esters between lipoproteins. Thus, variants in this gene have been associated with lower HDLc levels.90,91 Another interesting finding of this meta-analysis was the association of waist circumference-trigelycerides with a SNP located in GCKR (rs780093). The product of this gene is a regulatory protein that inhibits glucokinase activity in liver. Mutations in this gene could affect plasma glucose and triglyceride levels.92,93 The SNP rs301 located in the LPL gene is associated with the binary trait HDLc-waist circumference. LPL encodes a key enzyme in the metabolism of lipoproteins. This enzyme hydrolyzes triglycerides into free fatty acids and glycerol and converts very low-density lipoprotein to low density lipoprotein.94 The SNP rs10468017, a variant within LIPC gene, showed pleiotropic effects on waist circumference- HDLc. LIPC, like LPL, is a lipoprotein lipase but is expressed mainly in liver. Furthermore, it has been revealed as an important enzyme in HDLc metabolism.95

In the second meta-analysis of 2011, Kilpeläinen et al.31 meta-analyzed data from 45 studies of adults and confirmed that the SNP rs99396909, located in FTO gene, was associated with waist circumference at a genome-wide significance level. In addition, the authors performed a second meta-analysis to show the relationship between FTO-physical activity interaction and waist circumference. The association in both groups (physically active and physically inactive individuals) was genome-wide significance. Moreover, the results indicated that the influence of the FTO risk allele on waist circumference was 33% smaller in the active individuals than in the inactive ones.31

83 Chapter 1

Subsequent research by Fox et al.37 was based on a meta-analysis of GWAS for visceral adipose tissue and subcutaneous adipose tissue measured with CT. In relation to central obesity, 2 SNPs reached genome-wide significance level: LYPLAL1 (rs11118316) for the ratio of visceral adipose tissue to subcutaneous adipose tissue, previously in association with waist-hip ratio, and THNSL2 (rs1659258) for visceral adipose tissue in women only.32 However, the function of these genes in body fat distribution is still not clear.

Another 2-stage meta-analysis of GWAS was published in 2013 by Berndt et al.36 The first meta-analysis comprised a total of 51 studies to select SNPs for follow-up genotyping. Data from the stage 1 and follow-up studies were meta-analyzed. Four of the polymorphisms associated significantly with waist-hip ratio were previously described as waist-hip ratio-- associated loci in other populations: LY86 (rs1294421), RSPO3 (rs7745274), COBLL1 (rs13389219) and LYPLAL1 (rs2820464).32,35 COBLL1, a gene involved in neural tube formation, has also been associated with insulin resistance in adults and young individuals.96-98

Additionally, Randall et al.33 performed a sex-specific meta-analysis of 46 GWAS to determine sexually dimorphic associations with body fat distribution using waist circumference and waist- hip ratio. To identify and validate 348 SNPs derived from the stage-1 analyses, a meta-analysis of 48 additional studies was performed. The joint analysis of association showed 4 previously identified loci (near GERB14/COBLL1 [rs6717858], LYPLAL1/SLC30A10 [rs2820443], VEGFA [rs1358980] and ADAMTS9 [rs2371767]) and 3 novel loci (near MAP3K1 [rs11743303], HSD17B4 [rs10478424] and PPARG [rs4684854]), which were significant for waist circumference in women but not in men.32,36,37 MAP3K1 acts in the mitogen-activated protein kinase signaling pathway, which is activated by different intracellular and extracellular stimuli, including oxidative stress, cytokines and hormones, among others. Moreover, several polymorphisms of MAP3K are associated with triglyceride levels, cancer and neurodegenerative diseases.99,100 The protein encoded by HSD17B4 is a bifunctional enzyme (multifunctional enzyme type 2; MFE-2) that participates in the peroxisomal lipid metabolism.101 In this sense, peroxisomal oxidation of fatty acids produces oxidative stress, a process implicated in obesity. Interestingly, PPARG, which encodes a regulator of adipocyte differentiation (PPAR-gamma), is one of the genes most linked to obesity development and has also been implicated in lipid storage and insulin sensitization.102

The meta-analysis conducted by Liu et al.38 in populations of African ancestry further identified 2 loci influencing waist circumference and waist-hip ratio (rs2075064, in LHX2, and rs6931262, in RREB1, respectively). The protein encoded by LHX2 is a transcriptional regulator in central

84 Results nervous system and embryonic tissue development, including adipocyte differentiation.103 The function of RREB1 on adipose tissue has not been determined.

3.3. Central obesity and gene-diet interaction

Elucidating the mechanisms of the relationship between genetics and diet may provide a key to understanding the observed differences in genetic risk for central obesity in several populations, thereby facilitating personalized nutrition and therapy.104 Different studies have been carried out to achieve this aim (Table 3).

In this context, Song et al.45 analyzed the interaction between energy intake, a variant of the IL6 receptor (IL6R) gene (Asp358Ala or rs2228145) and waist circumference. The results demonstrated that, under excess energy intake, waist circumference was higher in TT and GT genotypes compared to non-T-allele carriers. The protein encoded by the IL6R belongs to the IL6 receptor complex. A dysregulated production of IL6, a proinflammatory cytokine, and its receptor have been shown to be associated with several metabolic disorders, autoimmune diseases and certain types of cancer.105-107

The amount and composition of dietary fat are two nutritional factors frequently investigated in relation to obesity and other metabolic disorders. The study published by Robitaille et al.43 demonstrated the potential relationship between a common genetic polymorphism of PPARG, P12A (rs1801282), and waist circumference, and whether this interaction was modulated by intake of dietary fat. The authors observed that the P12/P12 homozygotes with higher total fat and SFA intake, presented the higher waist circumference. In contrast, among the A12 allele carriers, waist circumference did not change with the amount of fat or SFA intake. A second study, by Dedoussis et al.,50 reported an association between PPARG P12A-diet interaction and central obesity in children. This research identified a gene-nutrient interaction with MUFA, whereby the P12/P12 homozygotes with higher intake of MUFA had lower waist circumference compared with those with lower intake. Considering the results of both studies, it can be concluded that homozygotes for the P allele who consumed a high-fat diet are at higher risk of central obesity, while the risk is lower when MUFA intake is high.

Another gene-diet interaction study related to central obesity was performed in 2007 by Robitaille et al.,44 who analyzed polymorphisms located in CPT1A (rs28936372) and CPT1B (rs470017) were analyzed. Regarding CPT1A, the A275/A275 homozygotes who consumed a high-fat diet had greater waist circumference than those who did not. Concerning CPT1B, the E531/K531 heterozygotes had a greater waist circumference when they consumed a high-fat

85 Chapter 1 diet. CPT1A and CPT1B, expressed in liver and heart/muscle respectively, are implicated in the mitochondrial oxidation of long-chain fatty acids.108

Philips et al.47 analyzed the cumulative risk of having several alleles located in STAT3, a gene involved in obesity-related traits. The risk alleles were defined as the G alleles for rs8069645, rs744166, rs1053005 and rs2293152. In individuals carrying more than 2 risk alleles, high SFA intake was associated with higher waist circumference. However, in individuals carrying 1 or fewer risk alleles, waist circumference was not affected by the amount of fat or SFA consumed. When the authors analyzed the interaction with MUFA and PUFA intakes, no differences were observed. The protein encoded by STAT3 is activated through phosphorylation in response to several cytokines and growth factors, which in turn are implicated in obesity.109 In fact, polymorphisms in STAT3 have been associated with body- weight regulation and glucose homeostasis.110

Furthermore, in 2010 Philips et al.48 aimed to determine the association between polymorphisms of IL-6, TNF-alpha and LTA and features of the metabolic syndrome as well as the interactions of the polymorphisms with total plasma fatty acids as a biomarker of habitual dietary fat intake. The results indicated that those with the risk genotype, defined as those carrying the LTA rs915654 A allele, the TNF-alpha rs1800629 GG genotype and the IL6 rs1800797 GG genotype, had higher waist circumferences than those without the risk genotype carriers when PUFA/SFA concentrations in plasma were low. Moreover, among those with the risk genotype, the highest PUFA/SFA levels were correlated with lower proinflammatory status, triacylglycerol levels, and the homeostasis model assessment-insulin resistance. TNF-alpha is implicated in a wide variety of functions, such as inflammation, cell proliferation, cell differentiation, cell apoptosis, lipid metabolism, and coagulation, in addition to disorders such as autoimmune diseases, insulin resistance and cancer.111 LTA, like IL6 and TNF-alpha, encodes for a cytokine that is involved in inflammatory, immunostimulatory, and antiviral responses as well as the formation of secondary lymphoid organs and apoptosis.112 Moreover, LTA is implicated in different metabolic disorders including insulin resistance, type 2 diabetes, and excess body weight.113

Another approach to the analysis of gene-diet interactions was used by Phillips et al.,49 who tested the effect of AAC2 gene polymorphisms. This study demonstrated that, when PUFA intake was high, waist circumference was greater in the AA homozygotes of the rs4766587 SNP compared with the GG homozygotes. ACC2 belongs to a complex multifunctional enzyme system that regulates the fatty acid synthesis and oxidation.114 Several studies have

86 Results demonstrated that ACC2 clearly plays a role in lipid metabolism, insulin resistance and obesity.115,116

In this context, Mattei et al.51 observed several interactions between SNPs in the cluster APOA1/C3/A4/A5 and intake of dietary fat. Among individuals with low fat intake, the homozygotes for the common allele of APOA1-75 (rs670) had lower waist circumferences than carriers of the minor allele. No differences between genetic variants were found among the high-fat consumers. Several polymorphisms of the APOA1 gene, which encodes the major protein component of HDLc in plasma, have been studied in relation to dyslipidemia, insulin resistance and metabolic syndrome in different populations.117,118

Philips et al.52 also investigated the relationship between the FTO rs9939609 polymorphism and waist circumference, and whether this association was modulated by fat intake. Carriers of the risk allele, AA and AT, showed significantly larger waist circumferences than TT homozygote individuals among high-SFA consumers at baseline and in the follow-up period. However, there was no difference by genotype within the high-SFA consumers. Moreover, in the carriers of the risk allele, there was an increase in waist circumference when the SFA intake was higher. Numerous studies have demonstrated that FTO rs9939609 is associated with a higher risk of obesity.5

Central obesity can also be modulated by carbohydrate intake. Smith et al.46 investigated the influence of complex carbohydrates on the effect of polymorphisms located in PLIN. Authors reported that, among low consumers of complex carbohydrates, A allele carriers (AA and AG) of the rs894160 SNP presented the highest waist circumference. In contrast, among high consumers of complex carbohydrates, carriers of the A allele had the lowest waist circumference. The protein encoded by PLIN plays a key role in adipocyte metabolism and lipolysis and in turn, influences obesity and insulin resistance.119,120 Polymorphisms located in PLIN have also been associated with other gene-diet interactions and weight loss.121

4. DNA METHYLATION AND OBESITY

A wide range of epigenetic changes in DNA over the lifetime may modify gene expression. The availability of microarray-based techniques has led to DNA methylation being the mechanism most studied in relation to nutrition and related diseases.17 Therefore, this review is focused on the studies that analyze DNA methylation in order to better understand the influence of lifestyle factors on the obesity phenotype.

87 Chapter 1

4.1. Epigenome-wide studies on obesity

EWAS allows collecting information about DNA methylation variations throughout the epigenome. Thus far, two EWAS have been carried out to characterize DNA methylation profile in obesity (Table 4).

Wang et al.53 used a case control study to compare obese male adolescents with lean male adolescents of African-American ancestry. Moreover, they performed a replication study in an independent sample. This investigation reported 11 significant CpG sites in the EWAS of DNA isolated from peripheral blood leukocytes. The CpG sites were located in or near novel genes that had not been previously related to obesity or metabolic-related traits. Interestingly, both Wang et al.53 and later Moleres et al.60 showed a differential methylation pattern in the HIPK3 gene. Wang et al.53 observed higher methylation levels in obese subjects, whereas Moleres et al.60 found that subjects with a high response to a weight loss program had significantly higher methylation levels in HIPK3 gene before the treatment. However, thus far there is no plausible information about the function of HIPK3 in metabolism-related pathways.

Feinberg et al.54 were interested in identifying variable methylated regions, which are defined by increased variability rather than an increase or decrease in DNA methylation, related to BMI using lymphocyte cells. Half of 227 variable methylated regions were identified by unbiased genome-scale analysis and appeared to be stable over time within individuals. Four of these variable methylated regions showed a relationship to obesity at 2 separate visits after 10 years: PM20D1, MMP9, PRKG1 and RFC5.

4.2. DNA methylation markers and body weight control

Most obesity treatments are based on dietary approaches.1 However, nutritional interventions aiming to reduce body weight have had varying degrees of success, depending on genetic factors.122 Thus, several studies have demonstrated that epigenetic markers could be involved in the response to weight-loss interventions (Table 5).

Campión et al.55 analyzed the methylation pattern of the TNF-alpha promoter in peripheral blood mononuclear cells before an energy-restriction intervention. After the intervention, the population was classified as high responders or low responders, according to body-weight loss. The analysis demonstrated that 2 CpG sites located in the TNF-alpha promoter were hypomethylated in the group of men with greater response to the dietary intervention. Hermsdorff et al.123 examined DNA isolated from white blood cells of normal-weight women

88 Results and reported that individuals with higher methylation levels of TNF-alpha promoter showed higher trunk fat.

Bouchard et al.56 investigated the differences in DNA methylation patters of adipose tissue between low and high responders (according to weight loss) to a hypocaloric diet. The microarray analysis identified 35 CpG sites and 3 CpG sites differentially methylated before and after the treatment, respectively. Four of these CpG sites showed significant methylation differences after the validation study by the Sequenom EpiTyper (Sequenom, San Diego, CA). Before the dietary intervention, 3 regions located in the DNASE1L2 gene were hypomethylated and 1 region located in the OAT gene was hypermthylated in the high responders compared with the low responders. DNASE1L2 is activated by 2 inflammatory cytokines, TNF-alpha and IL-1h, via the nuclear factor-nB (NFnB) pathway.124 Shiokawa et al.124 hypothesized that the proteins encoded by DNASE1L2 might play an important physiological role under inflammatory conditions, including obesity. The OAT gene, which encodes the mitochondrial enzyme ornithine aminotransferase, has not yet been associated with obesity or metabolic-related traits in humans.125

Using a study design similar to that of Bouchard et al.56, Milagro et al.57 identified, in peripheral blood mononuclear cells, different DNA methylation patterns of several CpG sites that were associated with the weight-loss outcome. A methylation microarray was performed to search for baseline and endpoint epigenetic differences between high and low responders to the intervention. A total of 432 CpG sites were differentially methylated before the weight-loss program and 15 CpG sites at the end of the intervention. After the validation process, 2 CpG sites located in the ATP10A gene were hypomethylated and 2 CpG sites in the CD44 gene were hypermethylated in the high-responder group at baseline. ATP10A encodes an aminophospholipid translocase protein that transports phosphatidylserine and phosphatidylethanolamine across the bilayer from side to side.126 Mice inheriting a maternal deletion of ATP10A develop obesity and type 2 diabetes.127 Recently, it has been shown in an animal model that CD44 plays an important role in regulating traits of metabolic syndrome, such as adipose inflammation, hepatic steatosis and insulin resistance.128 After calorie restriction, 9 CpG sites in the WT1 promoter were hypermethylated in the subjects who lost more body weight. This gene encodes a transcription factor that acts as an oncogene, mainly in Wilm’s tumor but also in other types of cancers, including leukaemia and breast cancer.129 Wilm’s tumor is accompanied in many cases by obesity and hypertension.130,131

In the same year (2011), Cordero et al.58 investigated the methylation profile of genes encoding adipokines and cytokines in subcutaneous adipose tissue in response to calorie

89 Chapter 1 restriction by using methylation-specific polymerase chain reaction. Before the weight-loss intervention, several CpG sites located in the promoter of leptin and TNF-alpha were slightly hypomethylated in the individuals who lost more weight. After the intervention, however, these differences were not observed. Previously, the same group had described that the methylation pattern of leptin was modulated by a high-fat diet in rats.132 The finding of slightly hypomethylated CpG sites in TNF-alpha is in accordance with those from Campión et al.,55 mentioned previously.

Clock genes, such as CLOCK and PER2, are implicated in the regulation of circadian metabolism and are also associated with weight-loss outcomes. Milagro et al.59 found, in DNA isolated from blood cells, that the baseline methylation of 1 CpG site located in the CLOCK gene and 2 CpG sites located in PER2 were directly associated with body-weight loss in women. Polymorphisms in CLOCK and PER have been previously associated with obesity and metabolic syndrome.133

A study by Moleres et al.60 reported new epigenetic biomarkers that could predict the result of a weight-loss intervention in adolescents. A validation process of the initial microarray indicated that specific regions located near or within AQP9, DUSP22, TNNT1, HIPK3, and TNNI3 were hypermethylated in the high-responder group compared with the low-responder group. Aquaporin 9 is a water-selective membrane channel that allows the passage of a wide variety of noncharged solutes, including glycerol.134 A recent review reported that the coordinating regulation of adipose AQP7 (another member of the family of water-selective membrane channel, which transports glycerol from the adipocytes) and hepatic AQP9 contributes to glycerol and glucose metabolism in vivo.135 DUSP22 acts as a negative regulator of the proinflammatory IL-6/LIF/STAT3-mediated signaling pathway by desphosphorylating STAT3, which is stimulated by leptin.136 The expression of TNNT1, which is involved in the regulation of striated muscle contraction, has been correlated with different compounds of lipid metabolism.137 Moreover, DNA methylation levels of TNNT1 were positively correlated with HDLc levels.138 Moleres et al.60 concluded that a higher methylation level of this gene might allow subjects to achieve greater metabolic benefits from a weight-loss program. The roles of HIPK3 and TNNI3 in obesity has not been yet elucidated.

DNA methylation biomarkers could be also useful fir predicting weight regain after a weight- loss intervention. Crujeiras et al.139 described, in leukocytes, the methylation levels of specific genes related to appetite according to the weight regain process after a dietary intervention. Thus, before the intervention, 1 CpG located in POMC was hypermethylated in regainers compared with non-regainers, whereas 3 CpG sites located in NPY were hypomethylated in

90 Results regainers. The authors hypothesized that the nonregainers could be partially protected against weight regain by an epigenetic mechanism involving the appetite regulatory neuropeptides.139 POMC has a functional role in the leptin-melanocortin pathway that controls human energy homeostasis, whereas NPY acts as a regulator of energy homeostasis and associated processes. 140,141

Several studies have investigated whether an obesogenic diet could influence DNA methylation. Jacobsen et al.142 performed a randomized crossover study comparing a high-fat overfeeding diet (60% fat, 32.5% carbohydrate and 7.5% protein) with a control diet (35% fat, 50% carbohydrate and 15% protein). The investigation concluded that a short-term, high-fat overfeeding diet can modify the methylation pattern of specific CpG sites located in the genes FAP, UGT2B7, GABRA3, GCNT4, APOH and FUT1 in human skeletal muscle. Moreover, the authors observed that DNA methylation changes induced by the high-fat overfeeding diet can be partly reversed after 6-8 weeks of wash-out period.

4.3. Methylation at early age and development of excess weight or adiposity

Pregnancy and early childhood are considered important epigenetic windows over a human’s lifetime. Thus, several investigations have studied the association between methylation levels at birth or during childhood and the development of overweight, obesity, and adiposity.

The first study in humans linking gene methylation at birth and its relationship with early-life obesity traits was published in 2011. Godfrey et al.143 examined 2 independent cohorts and showed that high methylation of a specific CpG site located in RXRA (measured at birth in umbilical cord tissue) was associated with higher adiposity in later childhood. Interestingly, the greater methylation of RXRA was also associated with a lower maternal carbohydrate intake during early pregnancy. The authors have proposed 2 potential mechanisms by which RXRA, a nuclear receptor that mediates the biological effects of retinoids, could be involved in obesity and adiposity.144 In the first one, RXRA induces transcription that is dependent on the PPARs, which in turn are involved in insulin sensitivity and body fat. In the other one, RXRA is located in a region that contains several positive regulatory elements of transcription.143 Moreover, a recent study has linked a polymorphism located in RXRA to lipid and lipoproteins levels.145

Relton et al.146 studied DNA isolated from cord blood that children who had higher BMI and fat mass at age 9 years had higher methylation levels in specific CpG sites located in different genes at birth than children who had lower BMI and fat mass. Specifically, methylation of CpG sites located in CASP10 were inversely associated with BMI; methylation levels in HLA-DOB and

91 Chapter 1

NID1 were negatively associated with fat mass; and methylation in CDKN1C and EPHA1 was positively associated with BMI and fat mass. The membrane protein encoded by NID1 is implicated in maintaining the structure of the extracellular matrix of adipocytes and can therefore play a role in adiposity.147 CDKN1C is considered as a negative regulator of cell proliferation and is usually upregulated in adipose tissue of obese individuals.148 EPHA1 belongs to the ephrin receptor subfamily, which is involved in regulating cell shape and cell movement mainly in the nervous system, and also plays a role in the modulation of insulin signaling, which in turn is involved in obesity.149,150 Thus far, there is not information about the possible roles of CASP10 and HLA-DOB in obesity or adiposity.

Perkins et al.151 observed that the methylation of 3 CpG sites located in the H19 gene in cord bloos leukocytes at birth was associated with excess body weight at approximately age 1 year of age. In humans, H19 (a non-coding RNA) and IGF2 (a member of the insulin family of polypeptide growth factors) are clustered in an imprinted region on chromosome 11. Genes involved in an imprinted region are expressed in a parent-of-origin-specific manner.152 In general, IGF2 promotes growth, whereas H19 inhibits growth. The maternal unmethylated allele at the IGF2/H19 locus activates IGF2 and silences H19, thus contributing to the inhibition of growth. Conversely, the paternal allele activates IGF2 and silences H19.153 According to these results, Huang et al.154 found a positive association between DNA methylation at the IGF2/H19 locus and subcutaneous adipose tissue measured by skin fold thickness and direct ultrasound at 17 years of age.

Perng et al.155 also conducted a study on DNA methylation and development of adiposity. They concluded that, in boys, a lower LINE-1 methylation (in peripheral leukocytes) at the time of the recruitment was associated with higher increase in adiposity (as measured by BMI, waist circumference and subscapular/triceps skin fold thickness ratio) after 3 years. LINE-1 is a retrotransposon widely expressed in the human genome and has been widely used for studying global methylation of the whole genome.156 It has been reported that LINE-1 may contribute to an increased risk of chronic diseases, such as cancer, cardiovascular disease, and diabetes.157-159

5. CONCLUSION

GWAS are providing valuable information to understand the genetics of several disorders, including obesity and body fat distribution. The first part of the present review highlights potential candidate genes associated with the development of central obesity. Many of the

92 Results polymorphisms identified in the studies reviewed are located in genes implicated in inflammation, lipid and glucose metabolism, and eating behavior.

The most interesting gene that has been associated with central obesity at a genome-wide level of significance is LYPLAL1. Five studies have shown specific SNPs located in this gene.32,33,35-37 Two SNPs, rs2820443 and rs48465678, were associated with waist-hip ratio after adjustment for BMI, indicating that both variants are specific to central obesity.33,35 Moreover, SNPs located in LYPLAL1 are associated with metabolic-related traits such as insulin resistance and triglyceride levels.77 This finding is in accordance with growing evidence that SNPs correlated with body fat distribution are also associated with metabolic disorders, whereas SNPs related to BMI in general do not usually show these associations.38 It is important to note that the relationship between LYPLAL1 and waist-hip ratio was sex dependent for 2 SNPs (rs2820443 and rs2605100) that were associated with waist-hip ratio only in women.33,35 In this context, it has been hypothesized that these sex-related differences could be explained by a complex interplay of genetic and hormonal factors.33

Another interesting candidate gene for the association between genetics and central obesity is VEGFA. SNPs in this gene were significantly associated with waist-hip ratio adjusted for BMI in 2 studies.32,33 As in the case of LYPLAL1, one of the SNPs, rs1358980, was associated with waist-hip ratio in women only.33 However, more studies are needed to explain the physiological mechanisms involved in this association. Although rs9939609 (FTO) is one of the SNPs most studied in relation to obesity, it has not been identified in many GWAS that examine central obesity. Nonetheless, several replication studies have found this polymorphism to be associated with measurements of central obesity, including waist circumference and waist-hip ratio.75,160

In general, GWAS and meta-analyses of GWAS have been performed mostly in adults of European ancestry. Therefore, more studies are needed in other periods of the life cycle (children, adolescents, and the elderly) and in populations from more regions around the world. Another consideration is that, in general, central obesity is analyzed by anthropometric measurements (waist circumference and waist-hip ratio). The results should be confirmed with other more complex measurements of central obesity, such as visceral adipose tissue measured by CT. In fact, no significant association between rs2605100 (located in the LYPLAL1 gene) and central obesity (measured by CT and dual energy X-ray absorptiometry) has been found in postmenopausal Japanese or Caucasian women (n=48).161

93 Chapter 1

Some findings have indicated that energy intake or specific macronutrients could modulate the genetic predisposition to abdominal obesity, suggesting that individuals with certain genotypes could benefited from dietary interventions. Few studies have analyzed the gene-diet interactions related to central obesity, and thus severall limitations should be mentioned. First, the sample size for this type of study must be much higher in order to identify reliable associations. Second, the studies were performed in specific populations. Therefore, since dietary habits and genetic background differ between populations of different regions, generalization of the findings to other populations should be made with caution. Third, dietary consumption is usually self-reported through the use of food frequency questionnaires or dietary records. As a result, inaccuracies in memory or deficiencies in food databases, among other factors, could contriute to bias. Lastly, similar to the interactions between genes and diet, other environmental factors may bepresent that could affect the results of the studies but were not taken into account, such as early-life events, smoking, or physical activity. Such factors, if examined, could modulate or interfere with the genetic data. Taking these limitations together, it is clear that additional studies are required.

Several studies have suggested that DNA methylation patterns could be involved in individual’s response to different dietary treatments and/or the development of obesity and other metabolic disorders. In fact, DNA methylation could be used to predict the success of a weight- loss intervention or the development of excess body weight or adiposity.162 However, since DNA methylation patterns are cell-specific and show high variability among individuals, there are some limitations to the application of epigenetic biomarkers in nutritional intervention and treatment. To overcome these limitations, more EWAS should be performed in larger populations to identify the most reliable epigenetic markers related to obesity in different cells and tissues (i.e, total blood, PBMCs, oral mucosa). The results of the EWAS should be confirmed in follow-up studies to determine the etiological role of disease-associated epigenetic variations.21 Moreover, studies of gene expression should be included and screened. Since epigenetic markers are dynamic and are dependent on environmental factors, epigenetic studies should control for factors such as age, early-life events, diet, diseases, physical activity, and toxic habits, among others. In addition, since the epigenetic markers are cell and tissue-specific, they should be correlated with blood or tissue cell counts (i.e, neutrophils, eosinophils). Although new technologies have been developed, there are millions of epigenetic variations (i.e, CpG sites) in the entire genom, and to date very few CpG sites have been studied in relation to body weight regulation. It is too soon to use epigenetic markers as biomarkers of obesity and other metabolic diseases, but they might be useul in

94 Results understanding, at least in part, the variability in body weight regulation and the interactions between the genetic background and lifestyle factors.

ACKNOWLEDGEMENTS

Funding and sponsorship. This work has been supported by Linea Especial “Nutrición, Obesidad y Salud” of the University of Navarra and CIBERobn/RETICS schedules (Instituto Carlos III). L.G. holds fellowship of Asociación de Amigos of University of Navarra.

Declaration of interest. The authors have no relevant interests to declare.

SUPPORTING INFORMATION

Additional Supporting Information may be found in the online version of this article at the publisher’s website:

Table S1 Single-nucleotide polymorphisms reported to be associated with central-obesity- related traits (waist circumference, waist-hip ratio, visceral adipose tissue, visceral adipose tissue to subcutaneous adipose tissue ratio) in genome-wide association studies with significance levels of P > 10−5.

Table S2 Single-nucleotide polymorphisms reported to be associated with central-obesity- related traits (waist circumference, waist-hip ratio, and visceral adipose tissue) in meta- analyses of genome-wide association studies; P values between P < 10−7 and P < 10−6.

95 Chapter 1

REFERENCES

1. Abete I, Astrup A, Martinez JA, et al. Obesity and the metabolic syndrome: role of different dietary macronutrient distribution patterns and specific nutritional components on weight loss and maintenance. Nutr Rev. 2010;68:214– 231. 2. Siddarth D. Risk factors for obesity in children and adults. J Investig Med. 2013;61:1039–1042. 3. Valassi E, Crespo I, Santos A, et al. Clinical consequences of Cushing’s syndrome. Pituitary. 2012;15:319–329. 4. Moore CJ. Social position, psychological stress, and obesity: a systematic review. J Acad Nutr Diet. 2012;112:518– 526. 5. Razquin C, Marti A, Martinez JA. Evidences on three relevant obesogenes: MC4R, FTO and PPARgamma. Approaches for personalized nutrition. Mol Nutr Food Res. 2011;55:136–149. 6. Hinnouho GM, Czernichow S, Dugravot A, et al. Metabolically healthy obesity and risk of mortality: does the definition of metabolic health matter? Diabetes Care. 2013;36:2294–2300. 7. Thomas EL, Frost G, Taylor-Robinson SD, et al. Excess body fat in obese and normal-weight subjects. Nutr Res Rev. 2012;25:150–161. 8. Manios Y, Kanellakis S, Moschonis G, et al. Estimation of abdominal fat mass: validity of abdominal bioelectrical impedance analysis and a newmodel based on anthropometry compared with dual-energy x-ray absorptiometry. Menopause. 2013;20:1280–1283. 9. Shuster A, Patlas M, Pinthus JH, et al. The clinical importance of visceral adiposity: a critical review of methods for visceral adipose tissue analysis. Br J Radiol. 2012;85:1–10. 10. Bush WS, Moore JH. Chapter 11: genome-wide association studies. PLoS Comput Biol. 2012;8:e1002822. doi: 10.1371/journal.pcbi.1002822. 11. Loos RJ. Genetic determinants of common obesity and their value in prediction. Best Pract Res Clin Endocrinol Metab. 2012;26:211–226. 12. McPherson JD, Marra M, Hillier L, et al. A physical map of the human genome. Nature. 2001;409:934–941. 13. International HapMap Consortium. The International HapMap project. Nature. 2003;426:789–796. 14. Day FR, Loos RJ. Developments in obesity genetics in the era of genome-wide association studies. J Nutrigenet Nutrigenomics. 2011;4:222–238. 15. Simopoulos AP. Nutrigenetics/nutrigenomics. Annu Rev Public Health. 2010;31:53–68. 16. San-Cristobal R, Milagro FI, Martinez JA. Future challenges and present ethical considerations in the use of personalized nutrition based on genetic advice. J Acad Nutr Diet. 2013;113:1447–1454. 17. Mansego ML, Milagro FI, Campion J, et al. Techniques of DNA methylation analysis with nutritional applications. J Nutrigenet Nutrigenomics. 2013;6:83– 96. 18. Milagro FI, Mansego ML, De Miguel C, et al. Dietary factors, epigenetic modifications and obesity outcomes: progresses and perspectives. Mol Aspects Med. 2013;34:782–812. 19. Ho JW, Bishop E, Karchenko PV, et al. ChIP-chip versus ChIP-seq: lessons for experimental design and data analysis. BMC Genomics 2011;12:134. doi: 10.1186/1471-2164-12-134. 20. Fazzari MJ, Greally JM. Introduction to epigenomics and epigenome-wide analysis. Methods Mol Biol. 2010;620:243–265. 21. Rakyan VK, Down TA, Balding DJ, et al. Epigenome-wide association studies for common human diseases. Nat Rev Genet. 2011;12:529–541. 22. Talens RP, Boomsma DI, Tobi EW, et al. Variation, patterns, and temporal stability of DNA methylation: considerations for epigenetic epidemiology. FASEB J. 2010;24:3135–3144.

96 Results

23. Kristiansson K, Perola M, Tikkanen E, et al. Genome-wide screen for metabolic syndrome susceptibility loci reveals strong lipid gene contribution but no evidence for common genetic basis for clustering of metabolic syndrome traits. Circ Cardiovasc Genet. 2012;5:242–249. 24. Wang K, Li WD, Zhang CK, et al. A genome-wide association study on obesity and obesity-related traits. PLoS One. 2011;6:e18939. doi: 10.1371/journal.pone.0018939. 25. Cho YS, Go MJ, Kim YJ, et al. A large-scale genome-wide association study of Asian populations uncovers genetic factors influencing eight quantitative traits. Nat Genet. 2009;41:527–534. 26. Chambers JC, Elliott P, Zabaneh D, et al. Common genetic variation near MC4R is associated with waist circumference and insulin resistance. Nat Genet. 2008;40:716–718. 27. Croteau-Chonka DC, Marvelle AF, Lange EM, et al. Genome-wide association study of anthropometric traits and evidence of interactions with age and study year in Filipino women. Obesity (Silver Spring). 2011;19:1019–1027. 28. Norris JM, Langefeld CD, Talbert ME, et al. Genome-wide association study and follow-up analysis of adiposity traits in Hispanic Americans: the IRAS Family Study. Obesity (Silver Spring). 2009;17:1932–1941. 29. Polasek O, Marusic A, Rotim K, et al. Genome-wide association study of anthropometric traits in Korcula Island, Croatia. Croat Med J. 2009;50:7–16. 30. Plourde M, Vohl MC, Bellis C, et al. A variant in the LRRFIP1 gene is associated with adiposity and inflammation. Obesity (Silver Spring). 2013;21:185–192. 31. Kilpeläinen TO, Qi L, Brage S, et al. Physical activity attenuates the influence of FTO variants on obesity risk: a meta-analysis of 218,166 adults and 19,268 children. PLoS Med. 2011;8:e1001116. doi: 10.1371/journal.pmed.1001116. 32. Heid IM, Jackson AU, Randall JC, et al. Meta-analysis identifies 13 new loci associated with waist-hip ratio and reveals sexual dimorphism in the genetic basis of fat distribution. Nat Genet. 2010;42:949–960. 33. Randall JC,Winkler TW, Kutalik Z, et al. Sex-stratified genome-wide association studies including 270,000 individuals show sexual dimorphism in genetic loci for anthropometric traits. PLoS Genet. 2013;9:e1003500. doi: 10.1371/journal.pgen.1003500. 34. Kraja AT, Vaidya D, Pankow JS, et al. A bivariate genome-wide approach to metabolic syndrome: STAMPEED Consortium. Diabetes. 2011;60:1329–1339. 35. Lindgren CM, Heid IM, Randall JC, et al. Genome-wide association scan metaanalysis identifies three loci influencing adiposity and fat distribution. PLoS Genet. 2009;5:e1000508. doi: 10.1371/journal.pgen.1000508. 36. Berndt SI, Gustafsson S, Magi R, et al. Genome-wide meta-analysis identifies 11 new loci for anthropometric traits and provides insights into genetic architecture. Nat Genet. 2013;45:501–512. 37. Fox CS, Liu Y, White CC, et al. Genome-wide association for abdominal subcutaneous and visceral adipose reveals a novel locus for visceral fat in women. PLoS Genet. 2012;8:e1002695. doi: 10.1371/journal.pgen.1002695. 38. Liu CT, Monda KL, Taylor KC, et al. Genome-wide association of body fat distribution in African ancestry populations suggests new loci. PLoS Genet. 2013;9:e1003681. doi: 10.1371/journal.pgen.1003681. 39. Heard-Costa NL, Zillikens MC, Monda KL, et al. NRXN3 is a novel locus for waist circumference: a genome-wide association study from the CHARGE Consortium. PLoS Genet. 2009;5:e1000539. doi: 10.1371/journal.pgen.1000539. 40. Hindorff LA, MacArthur J, Morales J, et al. A Catalog of Published Genome-Wide Association Studies. Updated August 8, 2014. Bethesda, MD: National Human Genome Research Institute. Available at: http://www.genome.gov/gwastudies/. Accessed September 11, 2013.

97 Chapter 1

41. Hoggart CJ, Clark TG, De IorioM, et al. Genome-wide significance for dense SNP and resequencing data. Genet Epidemiol. 2008;32:179–185. 42. Panagiotou OA, Ioannidis JP, Genome-Wide Significance Project. What should the genome-wide significance threshold be? Empirical replication of borderline genetic associations. Int J Epidemiol. 2012;41:273–286. 43. Robitaille J, Despres JP, Perusse L, et al. The PPAR-gamma P12A polymorphism modulates the relationship between dietary fat intake and components of the metabolic syndrome: results from the Québec Family Study. Clin Genet. 2003;63:109–116. 44. Robitaille J, Houde A, Lemieux S, et al. Variants within the muscle and liver isoforms of the carnitine palmitoyltransferase I (CPT1) gene interact with fat intake to modulate indices of obesity in French-Canadians. J Mol Med (Berl) 2007;85:129–137. 45. Song Y,Miyaki K, Araki J, et al. The interaction between the interleukin 6 receptor gene genotype and dietary energy intake on abdominal obesity in Japanese men. Metabolism. 2007;56:925–930. 46. Smith CE, Tucker KL, Yiannakouris N, et al. Perilipin polymorphism interacts with dietary carbohydrates to modulate anthropometric traits in Hispanics of Caribbean origin. J Nutr. 2008;138:1852–1858. 47. Phillips CM, Goumidi L, Bertrais S, et al. Dietary saturated fat modulates the association between STAT3 polymorphisms and abdominal obesity in adults. J Nutr. 2009;139:2011–2017. 48. Phillips CM, Goumidi L, Bertrais S, et al. Additive effect of polymorphisms in the IL-6, LTA, and TNF-alpha genes and plasma fatty acid level modulate risk for the metabolic syndrome and its components. J Clin Endocrinol Metab. 2010;95:1386–1394. 49. Phillips CM, Goumidi L, Bertrais S, et al. ACC2 gene polymorphisms, metabolic syndrome, and gene-nutrient interactions with dietary fat. J Lipid Res. 2010;51:3500–3507. 50. Dedoussis GV, Manios Y, Kourlaba G, et al. An age-dependent diet-modified effect of the PPARgamma Pro12Ala polymorphism in children. Metabolism. 2011;60:467–473. 51. Mattei J, Demissie S, Tucker KL, et al. The APOA1/C3/A4/A5 cluster and markers of allostatic load in the Boston Puerto Rican Health Study. Nutr Metab Cardiovasc Dis. 2011;21:862–870. 52. Phillips CM, Kesse-Guyot E, McManus R, et al. High dietary saturated fat intake accentuates obesity risk associated with the fat mass and obesity-associated gene in adults. J Nutr. 2012;142:824–831. 53. Wang X, Zhu H, Snieder H, et al. Obesity related methylation changes in DNA of peripheral blood leukocytes. BMC Med. 2010;8:87. doi: 10.1186/1741-7015-8-87. 54. Feinberg AP, Irizarry RA, Fradin D, et al. Personalized epigenomic signatures that are stable over time and covary with body mass index. Sci Transl Med. 2010;2:49ra67. doi: 10.1126/scitranslmed.3001262. 55. Campión J,Milagro FI, Goyenechea E, et al. TNF-alpha promoter methylation as a predictive biomarker for weight-loss response. Obesity (Silver Spring). 2009;17:1293–1297. 56. Bouchard L, Rabasa-Lhoret R, Faraj M, et al. Differential epigenomic and transcriptomic responses in subcutaneous adipose tissue between low and high responders to caloric restriction. Am J Clin Nutr. 2010;91:309– 320. 57. Milagro FI, Campion J, Cordero P, et al. A dual epigenomic approach for the search of obesity biomarkers: DNA methylation in relation to diet-induced weight loss. FASEB J. 2011;25:1378–1389. 58. Cordero P, Campion J, Milagro FI, et al. Leptin and TNF-alpha promoter methylation levels measured by MSP could predict the response to a low calorie diet. J Physiol Biochem. 2011;67:463–470.

98 Results

59. Milagro FI, Gomez-Abellan P, Campion J, et al. CLOCK, PER2 and BMAL1 DNA methylation: association with obesity and metabolic syndrome characteristics and monounsaturated fat intake. Chronobiol Int. 2012;29:1180– 1194. 60. Moleres A, Campion J, Milagro FI, et al. Differential DNA methylation patterns between high and low responders to a weight loss intervention in overweight or obese adolescents: the EVASYON study. FASEB J. 2013;27:2504–2512. 61. Fall T, Ingelsson E. Genome-wide association studies of obesity and metabolic syndrome. Mol Cell Endocrinol. 2012;382:740–757. 62. Santini F, Maffei M, Pelosini C, et al. Melanocortin-4 receptor mutations in obesity. Adv Clin Chem. 2009;48:95– 109. 63. Xi B, Cheng H, Shen Y, et al. Study of 11 BMI-associated loci identified in GWAS for associations with central obesity in the Chinese children. PLoS One. 2013;8:e56472. doi: 10.1371/journal.pone.0056472. 64. Garzon J, Rodriguez-Munoz M, de la Torre-Madrid E, et al. Effector antagonism by the regulators of G protein signalling (RGS) proteins causes desensitization ofmu-opioid receptors in the CNS. Psychopharmacology (Berl). 2005;180:1–11. 65. Hishimoto A, Liu QR, Drgon T, et al. Neurexin 3 polymorphisms are associated with alcohol dependence and altered expression of specific isoforms. Hum Mol Genet. 2007;16:2880–2891. 66. Docampo E, Ribases M, Gratacos M, et al. Association of neurexin 3 polymorphisms with smoking behavior. Genes Brain Behav. 2012;11:704–711. 67. Berridge KC, Ho CY, Richard JM, et al. The tempted brain eats: pleasure and desire circuits in obesity and eating disorders. Brain Res. 2010;1350:43–64. 68. Manderfield LJ, Daniels MA, Vanoye CG, et al. KCNE4 domains required for inhibition of KCNQ1. J Physiol. 2009;587:303–314. 69. van Vliet-Ostaptchouk JV, van Haeften TW, Landman GW, et al. Common variants in the type 2 diabetes KCNQ1 gene are associated with impairments in insulin secretion during hyperglycaemic glucose clamp. PLoS One. 2012;7:e32148. doi: 10.1371/journal.pone.0032148. 70. Smemo S, Tena JJ, Kim KH, et al. Obesity-associated variants within FTO form long-range functional connections with IRX3. Nature. 2014;507:371–375. 71. Speliotes EK,Willer CJ, Berndt SI, et al. Association analyses of 249,796 individuals reveal 18 new loci associated with body mass index. Nat Genet. 2010;42:937–948. 72. Scuteri A, Sanna S, Chen WM, et al. Genome-wide association scan shows genetic variants in the FTO gene are associated with obesity-related traits. PLoS Genet. 2007;3:e115. doi: 10.1371/journal.pgen.0030115. 73. Loos RJ, Bouchard C. FTO: the first gene contributing to common forms of human obesity. Obes Rev. 2008;9:246–250. 74. Liu C, Mou S, Pan C. The FTO gene rs9939609 polymorphism predicts risk of cardiovascular disease: a systematic review and meta-analysis. PLoS One. 2013;8:e71901. doi: 10.1371/journal.pone.007190. 75. Vasan SK, Karpe F, Gu HF, et al. FTO genetic variants and risk of obesity and type 2 diabetes: a meta-analysis of 28,394 Indians. Obesity (Silver Spring) 2013;22:964–970. 76. Blasius AL, Beutler B. Intracellular toll-like receptors. Immunity. 2010;32:305–315. 77. Bille DS, Banasik K, Justesen JM, et al. Implications of central obesity-related variants in LYPLAL1, NRXN3, MSRA, and TFAP2B on quantitative metabolic traits in adult Danes. PLoS One. 2011;6:e20640. doi: 10.1371/journal.pone.0020640.

99 Chapter 1

78. Maeda S, Tsukada S, Kanazawa A, et al. Genetic variations in the gene encoding TFAP2B are associated with type 2 diabetes mellitus. J Hum Genet. 2005;50:283–292. 79. Nordquist N, Gokturk C, Comasco E, et al. The transcription factor TFAP2B is associated with insulin resistance and adiposity in healthy adolescents. Obesity (Silver Spring). 2009;17:1762–1767. 80. Kim HY, Gladyshev VN. Methionine sulfoxide reductases: selenoprotein forms and roles in antioxidant protein repair in mammals. Biochem J. 2007;407:321–329. 81. Scherag A, Dina C, Hinney A, et al. Two new loci for body-weight regulation identified in a joint analysis of genome-wide association studies for early-onset extreme obesity in French and German study groups. PLoS Genet. 2010;6:e1000916. doi: 10.1371/journal.pgen.1000916. 82. Steinberg GR, Kemp BE, Watt MJ. Adipocyte triglyceride lipase expression in human obesity. Am J Physiol Endocrinol Metab. 2007;293:E958–E964. 83. Nishimura S,Manabe I, Nagasaki M, et al. Adipogenesis in obesity requires close interplay between differentiating adipocytes, stromal cells, and blood vessels. Diabetes. 2007;56:1517–1526. 84. Lijnen HR. Angiogenesis and obesity. Cardiovasc Res. 2008;78:286–293. 85. Schleinitz D, Kloting N, Lindgren CM, et al. Fat depot-specific mRNA expression of novel loci associated with waist-hip ratio. Int J Obes (Lond). 2014;38:120–125. 86. Gesta S, Bluher M, Yamamoto Y, et al. Evidence for a role of developmental genes in the origin of obesity and body fat distribution. Proc Natl Acad Sci USA. 2006;103:6676–6681. 87. Graff M, Fernandez-Rhodes L, Liu S, et al. Generalization of adiposity genetic loci to US Hispanic women. Nutr Diabetes 2013;3:e85. doi: 10.1038/nutd.2013.26. 88. Hotta K, Kitamoto A, Kitamoto T, et al. Replication study of 15 recently published loci for body fat distribution in the Japanese population. J Atheroscler Thromb. 2013;20:336–350. 89. Zheng XY, Zhao SP, Yan H. The role of apolipoprotein A5 in obesity and the metabolic syndrome. Biol Rev Camb Philos Soc. 2013;88:490–498. 90. Huggins GS, Papandonatos GD, Erar B, et al. Do genetic modifiers of high density lipoprotein cholesterol and triglyceride levels also modify their response to a lifestyle intervention in the setting of obesity and type-2 diabetes mellitus?: the Action for Health in Diabetes (Look AHEAD) study. Circ Cardiovasc Genet. 2013;6:391–399. 91. Agirbasli M, Eren F, Agirbasli D, et al. Multi-locus candidate gene analyses of lipid levels in a pediatric Turkish cohort: lessons learned on LPL, CETP, LIPC, ABCA1, and SHBG. OMICS. 2013;17:636–645. 92. Ling Y, Li X, Gu Q, et al. Associations of common polymorphisms in GCKR with type 2 diabetes and related traits in a Han Chinese population: a case-control study. BMC Med Genet. 2011;12:66. doi: 10.1186/1471-2350-12-66. 93. Varbo A, Benn M, Tybjaerg-Hansen A, et al. TRIB1 and GCKR polymorphisms, lipid levels, and risk of ischemic heart disease in the general population. Arterioscler Thromb Vasc Biol. 2011;31:451–457. 94. Ramasamy I. Recent advances in physiological lipoprotein metabolism. Clin Chem Lab Med. 2013;12:1–33. 95. Neale BM, Fagerness J, Reynolds R, et al. Genome-wide association study of advanced age-related macular degeneration identifies a role of the hepatic lipase gene (LIPC). Proc Natl Acad Sci U S A. 2010;107:7395–7400. 96. Carroll EA, Gerrelli D, Gasca S, et al. Cordon-bleu is a conserved gene involved in neural tube formation. Dev Biol. 2003;262:16–31. 97. Manning AK, Hivert MF, Scott RA, et al. A genome-wide approach accounting for body mass index identifies genetic variants influencing fasting glycemic traits and insulin resistance. Nat Genet. 2012;44:659–669. 98. Mancina RM, Burza MA, Maglio C, et al. The COBLL1 C allele is associated with lower seruminsulin levels and lower insulin resistance in overweight and obese children. Diabetes Metab Res Rev. 2013;29:413–416.

100 Results

99. Kim EK, Choi EJ. Pathological roles of MAPK signaling pathways in human diseases. Biochim Biophys Acta. 2010;1802:396–405. 100. Waterworth DM, Ricketts SL, Song K, et al. Genetic variants influencing circulating lipid levels and risk of coronary artery disease. Arterioscler Thromb Vasc Biol. 2010;30:2264–2276. 101. Mehtala ML, Lensink MF, Pietikainen LP, et al. On the molecular basis of D-bifunctional protein deficiency type III. PLoS One. 2013;8:e53688. 102. Galbete C, Toledo E, Martinez-Gonzalez MA, et al. Pro12Ala variant of the PPARG2 gene increases body mass index: an updated meta-analysis encompassing 49,092 subjects. Obesity (Silver Spring). 2013;21:1486–1495. 103. Dahl L, Richter K, Hagglund AC, et al. Lhx2 expression promotes self-renewal of a distinct multipotential hematopoietic progenitor cell in embryonic stem cellderived embryoid bodies. PLoS One. 2008;3:e2025. doi: 10.1371/journal.pone.0002025. 104. Marti A, Goyenechea E, Martinez JA. Nutrigenetics: a tool to provide personalized nutritional therapy to the obese. J Nutrigenet Nutrigenomics. 2010;3:157–169. 105. Abeywardena MY, Leifert WR, Warnes KE, et al. Cardiovascular biology of interleukin-6. Curr Pharm Des. 2009;15:1809–1821. 106. Hsieh CH, Hung YJ, Wu LI, et al. Interleukin-6 receptor gene 48892 A/C polymorphism is associated with metabolic syndrome in female Taiwanese adolescents. Genet Test Mol Biomarkers 2012;16:1376–1381. 107. Bondurant KL, Lundgreen A, Herrick JS, et al. Interleukin genes and associations with colon and rectal cancer risk and overall survival. Int J Cancer. 2013;132:905–915. 108. Bonnefont JP, Djouadi F, Prip-Buus C, et al. Carnitine palmitoyltransferases 1 and 2: biochemical, molecular and medical aspects. Mol Aspects Med. 2004;25:495–520. 109. Zhang SS, Liu MG, Kano A, et al. STAT3 activation in response to growth factors or cytokines participates in retina precursor proliferation. Exp Eye Res. 2005;81:103–115. 110. Wunderlich CM, Hovelmeyer N, Wunderlich FT. Mechanisms of chronic JAKSTAT3-SOCS3 signaling in obesity. JAKSTAT. 2013;2:e23878. doi: 10.4161/jkst.23878. 111. Tzanavari T, Giannogonas P, Karalis KP. TNF-alpha and obesity. Curr Dir Autoimmun. 2010;11:145–156. 112. Upadhyay V, Fu YX. Lymphotoxin signalling in immune homeostasis and the control of microorganisms. Nat Rev Immunol. 2013;13:270–279. 113. Mahajan A, Tabassum R, Chavali S, et al. Obesity-dependent association of TNFLTA locus with type 2 diabetes in North Indians. J Mol Med (Berl). 2010;88:515–522. 114. Wakil SJ, Abu-Elheiga LA. Fatty acid metabolism: target for metabolic syndrome. J Lipid Res. 2009;50(Suppl):S138–S143. 115. Riancho JA, Vazquez L, Garcia-Perez MA, et al. Association of ACACB polymorphisms with obesity and diabetes. Mol Genet Metab. 2011;104:670–676. 116. Glund S, Schoelch C, Thomas L, et al. Inhibition of acetyl-CoA carboxylase 2 enhances skeletal muscle fatty acid oxidation and improves whole-body glucose homeostasis in db/db mice. Diabetologia. 2012;55:2044–2053. 117. Zhong L, Li Q, Jiang Y, et al. The ApoB/ApoA1 ratio is associated withmetabolic syndrome and its components in a Chinese population. Inflammation. 2010;33:353–358. 118. Hsu MC, Chang CS, Lee KT, et al. Central obesity in males affected by a dyslipidemia-associated genetic polymorphism on APOA1/C3/A4/A5 gene cluster. Nutr Diabetes 2013;3:e61. doi: 10.1038/nutd.2013.2.

101 Chapter 1

119. Yu D, Li C, Xie J, et al. Association between three genetic variants of the perilipin gene (PLIN) and glucose metabolism: results from a replication study among Chinese adults and a meta-analysis. Endocr Res. 2013;38:263– 279. 120. Smith CE, Ordovas JM. Update on perilipin polymorphisms and obesity. Nutr Rev. 2012;70:611–621. 121. Corella D, Qi L, Tai ES, et al. Perilipin gene variation determines higher susceptibility to insulin resistance in Asian women when consuming a high-saturated fat, low-carbohydrate diet. Diabetes Care. 2006;29:1313–1319. 122. Hainer V, Zamrazilová H, Spálová J, et al. Role of hereditary factors in weight loss and its maintenance. Physiol Res. 2008;57(Suppl):S1–S15. 123. Hermsdorff HH, Mansego ML, Campion J, et al. TNF-alpha promoter methylation in peripheral white blood cells: relationship with circulating TNFalpha, truncal fat and n-6 PUFA intake in young women. Cytokine. 2013;64:265–271. 124. Shiokawa D, Matsushita T, Kobayashi T, et al. Characterization of the human DNAS1L2 gene and the molecular mechanism for its transcriptional activation induced by inflammatory cytokines. Genomics. 2004;84:95–105. 125. Renner AB, Walter A, Fiebig BS, et al. Gyrate atrophy: clinical and genetic findings in a female without arginine- restricted diet during her first 39 years of life and report of a new OAT gene mutation. Doc Ophthalmol. 2012;125:81–89. 126. Dhar M, Hauser L, Johnson D. An aminophospholipid translocase associated with body fat and type 2 diabetes phenotypes. Obes Res. 2002;10:695–702. 127. DuBose AJ, Johnstone KA, Smith EY, et al. Atp10a, a gene adjacent to the PWS/AS gene cluster, is not imprinted in mouse and is insensitive to the PWSIC. Neurogenetics. 2010;11:145–151. 128. Kang HS, Liao G, DeGraff LM, et al. CD44 plays a critical role in regulating diet-induced adipose inflammation, hepatic steatosis, and insulin resistance. PLoS One. 2013;8:e58417. doi: 10.1371/journal.pone.0058417. 129. Li X, Wang S, Sitaram RT, et al. Single nucleotide polymorphisms in the Wilms’ tumour gene 1 in clear cell renal cell carcinoma. PLoS One. 2013;8:e58396. doi:10.1371/journal.pone.0058396. 130. Fischbach BV, Trout KL, Lewis J, et al. WAGR syndrome: a clinical review of 54 cases. Pediatrics. 2005;116:984– 988. 131. Steege A, Fahling M, Paliege A, et al. Wilms’ tumor protein (-KTS) modulates renin gene transcription. Kidney Int. 2008;74:458–466. 132. Milagro FI, Campion J, Garcia-Diaz DF, et al. High fat diet-induced obesitymodifies the methylation pattern of leptin promoter in rats. J Physiol Biochem. 2009;65:1–9. 133. Gomez-Abellan P, Madrid JA, Ordovas JM, et al. Chronobiological aspects of obesity and metabolic syndrome. Endocrinol Nutr. 2012;59:50–61. 134. Maeda N. Implications of aquaglyceroporins 7 and 9 in glycerol metabolism and metabolic syndrome. Mol Aspects Med. 2012;33:665–675. 135. Maeda N, Funahashi T, Shimomura I. Cardiovascular-metabolic impact of adiponectin and aquaporin. Endocr J. 2013;60:251–259. 136. Sekine Y, Tsuji S, Ikeda O, et al. Regulation of STAT3-mediated signaling by LMW-DSP2. Oncogene. 2006;25:5801–5806. 137. Nayak SB, Pinto Pereira LM, Boodoo S, et al. Association of troponin T and altered lipid profile in patients admitted with acute myocardial infarction. Arch Physiol Biochem. 2010;116:21–27. 138. Guay SP, Voisin G, Brisson D, et al. Epigenome-wide analysis in familial hypercholesterolemia identified new loci associated with high-density lipoprotein cholesterol concentration. Epigenomics. 2012;4:623–639.

102 Results

139. Crujeiras AB, Campion J, Diaz-Lagares A, et al. Association ofweight regain with specific methylation levels in the NPY and POMC promoters in leukocytes of obese men: a translational study. Regul Pept. 2013;186:1–6. 140. Beckers S, Zegers D, Van Gaal LF, et al. The role of the leptin-melanocortin signalling pathway in the control of food intake. Crit Rev Eukaryot Gene Expr. 2009;19:267–287. 141. Zhang L, Bijker MS, Herzog H. The neuropeptide Y system: pathophysiological and therapeutic implications in obesity and cancer. Pharmacol Ther. 2011;131:91–113. 142. Jacobsen SC, Brons C, Bork-Jensen J, et al. Effects of short-term high-fat overfeeding on genome-wide DNA methylation in the skeletal muscle of healthy young men. Diabetologia. 2012;55:3341–3349. 143. Godfrey KM, Sheppard A, Gluckman PD, et al. Epigenetic gene promoter methylation at birth is associated with child’s later adiposity. Diabetes. 2011;60:1528–1534. 144. Dolle P. Developmental expression of retinoic acid receptors (RARs). Nucl Recept Signal. 2009;7:e006. doi: 10.1621/nrs.07006. 145. Lima LO, Almeida S, Hutz MH, et al. PPARA, RXRA, NR1/2 and NR1/3 gene polymorphisms and lipid and lipoprotein levels in a Southern Brazilian population. Mol Biol Rep. 2013;40:1241–1247. 146. Relton CL, Groom A, St Pourcain B, et al. DNA methylation patterns in cord blood DNA and body size in childhood. PLoS One. 2012;7:e31821. doi: 10.1371/journal.pone.0031821. 147. Mariman EC, Wang P. Adipocyte extracellular matrix composition, dynamics and role in obesity. Cell Mol Life Sci. 2010;67:1277–1292. 148. Qiu J, Ni YH, Gong HX, et al. Identification of differentially expressed genes in omental adipose tissues of obese patients by suppression subtractive hybridization. Biochem Biophys Res Commun. 2007;352:469–478. 149. Duffy SL, Coulthard MG, Spanevello MD, et al. Generation and characterization of EphA1 receptor tyrosine kinase reporter knockout mice. Genesis. 2008;46:553–561. 150. Pasquale EB. Eph-ephrin bidirectional signaling in physiology and disease. Cell. 2008;133:38–52. 151. Perkins E, Murphy SK, Murtha AP, et al. Insulin-like growth factor 2/H19 methylation at birth and risk of overweight and obesity in children. J Pediatr. 2012;161:31–39. 152. Renfree MB, Suzuki S, Kaneko-Ishino T. The origin and evolution of genomic imprinting and viviparity in mammals. Philos Trans R Soc Lond B Biol Sci. 2013;368:20120151. doi: 10.1098/rstb.2012.0151. 153. Edwards CA, Ferguson-Smith AC. Mechanisms regulating imprinted genes in clusters. Curr Opin Cell Biol. 2007;19:281–289. 154. Huang RC, Galati JC, Burrows S, et al. DNA methylation of the IGF2/H19 imprinting control region and adiposity distribution in young adults. Clin Epigenetics 2012;4:21. doi:10.1186/1868-7083-4-21. 155. Perng W, Mora-Plazas M, Marin C, et al. A prospective study of LINE-1 DNA methylation and development of adiposity in school-age children. PLoS One. 2013;8:e62587. doi: 10.1371/journal.pone.0062587. 156. Zhu ZZ, Hou L, Bollati V, et al. Predictors of global methylation levels in blood DNA of healthy subjects: a combined analysis. Int J Epidemiol. 2012;41:126–139. 157. Murata A, Baba Y, Watanabe M, et al. Methylation levels of LINE-1 in primary lesion and matched metastatic lesions of colorectal cancer. Br J Cancer. 2013;109:408–415. 158. Cash HL, McGarvey ST, Houseman EA, et al. Cardiovascular disease risk factors and DNA methylation at the LINE-1 repeat region in peripheral blood from Samoan Islanders. Epigenetics. 2011;6:1257–1264. 159. Pearce MS, McConnell JC, Potter C, et al. Global LINE-1 DNA methylation is associated with blood glycaemic and lipid profiles. Int J Epidemiol. 2012;41:210–217.

103 Chapter 1

160. Vasan SK, Fall T, Job V, et al. A common variant in the FTO locus is associated with waist-hip ratio in Indian adolescents. Pediatr Obes. 2013;8:e45–e49. 161. Lim U, Ernst T, Wilkens LR, et al. Susceptibility variants for waist size in relation to abdominal, visceral, and hepatic adiposity in postmenopausal women. J Acad Nutr Diet. 2012;112:1048–1055. 162. Martinez JA, Milagro FI, Claycombe KJ, et al. Epigenetics in adipose tissue, obesity, weight loss and diabetes. Adv Nutr. 2014;5:71–81.

104 Results

Table 1 Single-nucleotide polymorphisms reported to be associated to central-obesity-related traits (waist circumference, waist-hip ratio, visceral adipose tissue, visceral adipose tissue to subcutaneous adipose tissue ratio) in genome-wide association studies reaching significance levels of p< 10-6

GWAS Other analysis Reference Gene SNP Population Sex Trait No. of subjects p value No. of subjects P value

FTO rs9940128 Finnish F and M WC 11,616 1.66 x 10^-9 - - Kristiansson et al. (2012)23 NRXN3 rs11624704 Non-Hispanic Caucasians (European ancestry) F and M WHR 1,060 2.67 x 10^-9 - - Wang et al. (2011)24 NCAM2 rs11088859 Non-Hispanic Caucasians (European ancestry) F and M WC 1,060 3.75 x 10^-8 - - Wang et al. (2011)24 C12orf51 rs2074356 East Asian F and M WHR 8,842 1.8 x 10^-7 16,703 7.8 x 10^-12a Cho et al. (2009)25 INHBB rs7581710 Non-Hispanic Caucasians (European ancestry) F and M WHR 1,060 2.12 x 10^-7 - - Wang et al. (2011)24 ARG1 rs2807278 Non-Hispanic Caucasians (European ancestry) F and M WHR 1,060 3.18 x 10^-7 - - Wang et al. (2011)24 TCBA1 rs870583 Indian Asian, European ancestry F and M WHR 2,684 4.5 x 10^-7 14,639 0.001b Chambers et al. (2008)26 KCNE4 rs1440072 Filipino F WC 1,792 7.87 x 10^-7 - - Croteau-Chonka et al. (2011)27 - rs4134351 Hispanic ancestry (Hispanic-Americans) F and M VAT/SAT (CT) 229 1.6 x 10^-6 1,190 1.9 x 10^-4c Norris et al. (2010)28 SEZ6L2 rs4787483 Croatia F and M WC 898 2.10 x 10^-6 - - Polasek et al. (2009)29 ANKS1B rs2373011 Filipino F WC 1,792 2.46 x 10^-6 - - Croteau-Chonka et al. (2011)27 IL18 rs2043055 Indian Asian, European ancestry F and M WHR 2,684 3.4 x 10^-6 14,639 0.02b Chambers et al. (2008)26 LRRFIP1 rs11680012 French-Canadian families F and M VAT (CT) 926 3.5 x 10^-6 - - Plourde et al. (2013)30 MAX rs7158173 Croatia F and M WC 898 3.93 x 10^-6 - - Polasek et al. (2009)29 IL1RAP rs9290936 Filipino F WC 1,792 4.03 x 10^-6 - - Croteau-Chonka et al. (2011)27 TNIK rs9811628 Indian Asian, European ancestry F and M WHR 2,684 4.1 x 10^-6 14,639 0.008b Chambers et al. (2008)26 CSNK1A1 rs353243 Indian Asian, European ancestry F and M WHR 2,684 4.1 x 10^-6 14,639 0.006b Chambers et al. (2008)26 KLHL36 rs11647936 Filipino F WC 1,792 4.16 x 10^-6 - - Croteau-Chonka et al. (2011)27 - rs7543757 Hispanic ancestry (Hispanic-Americans) F and M VAT (CT) 229 4.6 x 10^-6 1,190 9.1 x 10^-3c Norris et al. (2009)28 MC4R rs12970134 Indian Asian, European ancestry F and M WC 2,684 4.6 x 10^-6 14,639 1.7 x 10^-9b Chambers et al. (2008)26 CCDC99 rs13156607 Filipino F WC 1,792 4.61 x 10^-6 - - Croteau-Chonka et al. (2011)27 PPM1H rs7302017 Filipino F WC 1,792 5.47 x 10^-6 - - Croteau-Chonka et al. (2011)27 NDUFA5 rs805786 Indian Asian, European ancestry F and M WC 2,684 5.5 x 10^-6 14,639 0.11b Chambers et al. (2008)26 NDUFA5 rs805785 Indian Asian, European ancestry F and M WC 2,684 5.7 x 10^-6 14,639 0.15b Chambers et al. (2008)26 N/A rs9313296 Filipino F WC 1,792 5.72 x 10^-6 - - Croteau-Chonka et al. (2011)27 - rs4541696 Hispanic ancestry (Hispanic-Americans) F and M VAT/SAT (CT) 229 6.1 x 10^-6 1,190 3.6 x 10^-3b Norris et al. (2010)28 - rs17089410 East Asian F and M WHR 8,842 6.1 x 10^-6 16,703 4.4 x 10^-3a Cho et al. (2009)25 MC4R rs4450508 Indian Asian, European ancestry F and M WC 2,684 6.5 x 10^-6 14,639 1.8 x 10^-8b Chambers et al. (2008)26 MC4R rs502933 Indian Asian, European ancestry F and M WC 2,684 6.6 x 10^-6 14,639 4.4 x 10^-7b Chambers et al. (2008)26

105 Chapter 1

LRRC16 rs301396 Indian Asian, European ancestry F and M WC 2,684 6.7 x 10^-6 14,639 0.32b Chambers et al. (2008)26 MC4R rs477181 Indian Asian, European ancestry F and M WC 2,684 7.0 x 10^-6 14,639 8.1 x 10^-7b Chambers et al. (2008)26 SQRDL rs12594515 Filipino F WC 1,792 7.16 x 10^-6 - - Croteau-Chonka et al. (2011)27 GNA14 rs4745641 Indian Asian, European ancestry F and M WC 2,684 8.6 x 10^-6 14,639 0.1b Chambers et al. (2008)26 IL1RAP rs3773996 Filipino F WC 1,792 9.96 x 10^-6 - - Croteau-Chonka et al. (2011)27 Abbreviations: CT, computed tomography; F, female; GWAS, genome-wide association study; M, male; N/A, no protein-coding gene within 500 kb of the SNP; SNP, single nucleotide polymorphism; VAT, visceral adipose tissue; VAT/SAT, visceral adipose tissue to subcutaneous adipose tissue ratio; WC, Waist circumference; WHR, waist-hip ratio; -, no data available. aThis research performed a two stage GWAS. Data presented correspond to a combined analysis of stage 1 and stage 2. bThis research performed a two stage GWAS. Data presented correspond to a combined analysis of stage 1 and stage 2. The sample of stage 1 was Indian-Asian men, whereas the sample of the stage 2 and combined analysis was Indian-Asian and European ancestry. cThis research performed a three stage GWAS. Data presented correspond to a replication study (stage2). Data are sorted by GWAS P value.

106 Results

Table 2 Single-nucleotide polymorphisms reported to be associated with central-obesity-related traits (waist circumference, waist-hip ratio, and visceral adipose tissue) in meta-analyses of genome-wide association studies reaching significance levels of P< 10-8

Discovery meta-analysis Follow-up studies Combined meta-analysis Reference Gene SNP Population Sex Trait No. of subjects P value No. of subjects P value No. of subjects P value

Whites, Asians, African FTO rs9939609 F and M WC 159,848 5.4 x 10^-43 - - - - Kilpeläinen et al. (2011)31 Americans, Hispanics RSPO3 rs9491696 European ancestry F and M WHR adj BMI 77,164 2.10 x 10^-14 113,582 3.27 x 10^-28 190,746 1.84 x 10^-40 Heid et al. (2010)32 European, European LYPLAL1/SLC30A10 rs2820443 F WHR adj BMI 73,137 3.69 x 10^-18 74,657 9.51 x 10^-37 147,794 4.62 x 10^-37 Randall et al. (2013)33 American descent European, European VEGFA rs1358980 F WHR adj BMI 73,137 1.11 x 10^-13 74,657 1.38 x 10^-19 147,794 2.41 x 10^-31 Randall et al. (2013)33 American descent European, European GRB14/COBLL1 rs6717858 F WHR adj BMI 73,137 2.78 x 10^-15 74,657 3.64 x 10^-16 147,794 1.99 x 10^-29 Randall et al. (2013)33 American descent VEGFA rs6905288 European ancestry F and M WHR adj BMI 77,129 4.72 x 10^-10 95,430 1.18 x 10^-16 172,559 5.88 x 10^-25 Heid et al. (2010)32 TBX15-WARS2 rs984222 European ancestry F and M WHR adj BMI 77,167 3.81 x 10^-14 109,623 1.56 x 10^-12 186,790 8.69 x 10^-25 Heid et al. (2010)32 NFE2L3 rs1055144 European ancestry F and M WHR adj BMI 77,145 1.49 x 10^-8 113,636 3.26 x 10^-18 190,781 9.97 x 10^-25 Heid et al. (2010)32 GRB14 rs10195252 European ancestry F and M WHR adj BMI 77,119 3.23 x 10^-10 102,449 3.18 x 10^-16 179,568 2.09 X 10^-24 Heid et al. (2010)32 European, European ADAMTS9 rs2371767 F WHR adj BMI 73,137 1.63 x 10^-8 74,657 8.55 x 10^-17 147,794 7.07 x 10^-23 Randall et al. (2013)33 American descent LYPLAL1 rs4846567 European ancestry F and M WHR adj BMI 77,167 2.37 x 10^-12 91,820 3.15 x 10^-10 168,987 6.89 x 10^-21 Heid et al. (2010)32 DNM3-PIGC rs1011731 European ancestry F and M WHR adj BMI 77,094 1.72 x 10^-10 92,018 7.47 x 10^-9 169,112 9.51 x 10^-18 Heid et al. (2010)32 ITPR2-SSPN rs718314 European ancestry F and M WHR adj BMI 77,167 2.41 x 10^-8 107,503 1.49 x 10^-10 184,670 1.14 x 10^-17 Heid et al. (2010)32 LY86 rs1294421 European ancestry F and M WHR adj BMI 77,154 6.31 x 10^-9 102,189 2.69 x 10^-10 179,343 1.75 x 10^-17 Heid et al. (2010)32 HOXC13 rs1443512 European ancestry F and M WHR adj BMI 77,165 3.33 x 10^-8 112,353 2.92 x 10^-10 189,518 6.38 x 10^-17 Heid et al. (2010)32 CETP rs173539 European ancestry F and M HDLc-WC 22,161 1.0 x 10^-16 - - - - Kraja et al. (2011)34 ZNF259 rs2075290 European ancestry F and M WC-TG 22,161 1.1 x 10^-16 - - - - Kraja et al. (2011)34 APOA5 rs2266788 European ancestry F and M WC-TG 22,161 2.2 x 10^-16 - - - - Kraja et al. (2011)34 BUD13 rs10790162 European ancestry F and M WC-TG 22,161 6.6. x 10^-16 - - - - Kraja et al. (2011)34 European, European PPARG rs4684854 F WHR adj BMI 73,137 2.36 x 10^-8 74,657 1.48 x 10^-7 147,794 4.17 x 10^-14 Randall et al. (2013)33 American descent ADAMTS9 rs6795735 European ancestry F and M WHR adj BMI 77,162 2.47 x 10^-7 84,480 6.75 x 10^-8 161,642 9.79 x 10^-14 Heid et al. (2010)32 GCKR rs780093 European ancestry F and M WC-TG 22,161 1.9 x 10^-12 - - - - Kraja et al. (2011)34 ZNRF3-KREMEN1 rs4823006 European ancestry F and M WHR adj BMI 77,086 2.41 x 10^-5 93,911 2.41 x 10^-5 170,997 1.10 x 10^-11 Heid et al. (2010)32 2.57 x 20^-1 - TFAP2B rs987237 European ancestry F and M WC 38,635 1.10 x 10^-4 12,369-43,016 118,691 1.87 x 10^-11 Lindgren et al. (2009)35 2.22 x 10^-5 European, European MAP3K1 rs11743303 F WHR adj BMI 73,137 2.27 x 10^-6 74,657 7.15 x 10^-7 147,794 2.69 x 10^-11 Randall et al. (2013)33 American descent LPL rs301 European ancestry F and M HDLc-WC 22,161 3.2 x 10^-11 - - - - Kraja et al. (2011)34 LY86 rs1294421 European ancestry F and M WHR 9,803 1.62 x 10^-9 6,703 6.85 x 10^-3 16,506 2.19 x 10^-10 Berndt et al. (2013)36 NISCH-STAB1 rs6784615 European ancestry F and M WHR adj BMI 76,859 3.18 x 10^-7 109,028 1.56 x 10^4 185,887 3.84 x 10^-10 Heid et al. (2010)32

107 Chapter 1

RSPO3 rs7745274 European ancestry F and M WHR 10,077 3.91 x 10^-6 6,250 5.33 x 10^-5 16,327 8.80 x 10^-10 Berndt et al. (2013)36 CCDC121 rs3749147 European ancestry F and M WC-TG 22,161 1.4 x 10^-9 - - - - Kraja et al. (2011)34 CPEB4 rs6861681 European ancestry F and M WHR adj BMI 77,164 1.40 x 10^-6 85,722 2.13 x 10^-4 162,886 1.91 x 10^-9 Heid et al. (2010)32 C2orf16 rs1919128 European ancestry F and M WC-TG 22,161 2.0 x 10^-9 - - - - Kraja et al. (2011)34 European, European HSD17B4 rs10478424 F WHR adj BMI 73,137 1.02 x 10^-5 74,657 3.81 x 10^-5 147,794 3.45 x 10^-9 Randall et al. (2013)33 American descent ZNF512 rs13022873 European ancestry F and M WC-TG 22,161 5.0 x 10^-9 - - - - Kraja et al. (2011)34 LYPLAL1 rs2820464 European ancestry F and M WHR 9,458 4.54 x 10^-10 6,177 1.16 x 10^-1 15,635 7.35 x 10^-9 Berndt et al. (2013)36 4.63 x 10^-1 - MSRA rs545854 European ancestry F and M WC 36,865 1.32 x 10^-5 3,406-31,481 80,210 8.89 x 10^-9 Lindgren et al. (2009)35 5.31 x 10^-4 Caucasians, African THNSL2 rs1659258 F VAT (CT) 10,557 1.58 x 10^-8 3,158 1.6 x 10^-8 - - Fox et al. (2012)37 Americans LHX2 rs2075064 African ancestry F and M WC adj BMI 23,564 5.5 x 10^-8 10,027 3.2 x 10^-2 33,591 2.2 x 10^-8 Liu et al. (2013)38 RREB1 rs6931262 African ancestry F and M WHR adj BMI 19,744 5.3 x 10^-8 7,606 4.5 x 10^-2 27,350 2.5 x 10^-8 Liu et al. (2013)38 8.17 x 10^-2 - LYPLAL1 rs2605100 European ancestry F WHR 21,397 1.30 x 10^-8 6,021-20,213 47,633 2.55 x 10^-8 Lindgren et al. (2009)35 9.06 x 10^-2 COBLL1 rs13389219 European ancestry F and M WHR 9,880 2.68 x 10^-6 6,772 3.09 x 10^-3 16,652 3.24 x 10^-8 Berndt et al. (2013)36 LOC100129150 rs9987289 European ancestry F and M HDLc-WC 22,161 3.7 x 10^-8 - - - - Kraja et al. (2011)34 NRXN3 rs10146997 Caucasian descent F and M WC 31,373 6.4 x 10^-7 38,641 0.009 70,014 5.3 x 10^-8 Heard-Costa et al. (2009)39 LIPC rs10468017 European ancestry F and M HDLc-WC 22,161 5.5 x 10^-8 - - - - Kraja et al. (2011)34 Abbreviations: adj, adjusted for; BMI, body mass index; CT, computed tomography; F, female; HDLc, high density lipoprotein cholesterol; M, male; SNP, Single nucleotide polymorphism; VAT, Visceral adipose tissue; TG, triglycerides WC, waist circumference; WHR, waist-hip ratio; -n no available data. Data were sorted first by combined meta-analysis P values, then by follow up study P values, and lastly by discovery meta-analysis P values.

108 Results

Table 3 Single-nucleotide polymorphism-diet interactions in relation to central-obesity-related traits (waist circumference)

No.of Reference Gene SNP Population Sex Energy or nutrient Results subjects Robitaille et al. The higher TF intake, the greater WC in P12/P12 PPARG rs1801282 French-Canadian F and M 720 TF and SFA (2003)43 homozygotes Robitaille et al. The higher TF intake, the greater WC in A275/A275 CPT1A rs17610395 French-Canadian F and M 351 TF (2007)44 homozygotes Robitaille et al. The higher TF intake, the greater WC in E531/K531 CPT1B rs470017 French-Canadian F and M 351 TF (2007)44 heterozygotes Song et al. The higher energy intake, the greater WC in T allele carriers 45 IL6R rs8192284 Japanese M 285 Energy (2007) (TT and TG) Smith et al. The lower complex CH intake, the greater WC in A allele PLIN rs894160 Puerto Rican F and M 920 Complex CH (2008)46 carriers Smith et al. The higher complex CH intake, the lower WC in A allele PLIN rs894160 Puerto Rican F and M 921 Complex CH (2008)46 carriers rs8069645, rs744166, Phillips et al. The higher SFA intake, the greater WC in subjects carrying STAT3 rs1053005, rs2293152, French F and M 1754 SFA (2009)47 >2 risk alleles rs2306580 Philips et al. IL6, LTA, TNF- rs1800797, rs1800629, SFA/PUFA plasma The lower PUFA/SFA plasma concentrations, the greater WC 48 French F and M 1754 (2010) alpha rs915654 concentrations in risk genotype carriers Philips et al. ACC2 rs4766587 French F and M 464 PUFA The higher PUFA intake, the greater WC in AA homozygotes (2010)49 Dedoussis et al. The higher MUFA intake, the lower WC in P12/P12 PPARG rs1801282 Greek (young) F 1332 MUFA (2011)50 homozygotes Mattei et al. The lower TF intake, the lower WC in common allele APOA1 rs670 Puerto Rican F and M 821 TF (2011)51 homozygotes Philips et al. FTO rs9939609 French F and M 1754 SFA The higher SFA intake, the greater WC in A allele carriers (2012)52 Abbreviations: CH, carbohydrates; F, female; M, male; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids; SFA/PUFA, saturated fatty acids to polyunsaturated fatty acids ratio; SNP, single nucleotide polymorphism; TF, total fat; WC, waist circumference

109 Chapter 1

Table 4 Epigenome-wide association studies CpG sites associated with obesity

Reference Population No. of subjects Tissue/Cell Gene Chr CpG location P value Wang et al. Males of 14 Leukocyte UBASH3A 21 -156a 5.00 x 10^-6 53 (2010) African- a American CTNND1 11 -431 2.00 x 10^-5 ancestry TRIM3 11 -348a 2.40 x 10^-5 CTSZ 20 -292a 4.40 x 10^-5 TRAF5 1 -319a 4.50 x 10^-5 HIPK3 11 +390a 6.50 x 10^-5 DPCR1 6 -46a 6.90 x 10^-5 HIF3A 19 +153a 1.08 x 10^-4 NOTCH4 6 -51a 1.15 x 10^-4 CDH5 16 -243a 1.75 x 10^-4 CREB3L3 19 +42a 1.78 x 10^-3 Feinberg et Icelandic 64 Lymphocyte PM20D1b 1 - 3.32 x 10^-3 al. (2010)54 population MMP9b 20 2644c 6.58 x 10^-3 PRKG1b 10 1096c 1.32 x 10^-2 RFC5b 12 49957c 1.75 x 10^-2 Abbreviations: Chr, chromosome. a Distance to the transcription start site. b Variably methylated regions (VMRs). c Distance from the nearest gene.

110 Results

Table 5 DNA methylation markers before and after the weight-loss intervention

Reference Population No. of subjects Tissue/cell Other details Gene Chr CpG location P value Biomarkers before the weight loss intervention Campión et al. (2009)55 Obese men 12 PBMCs 8-week WLI/ high responders: lost ≥5% W TNF-alpha total methylation 6 - 0.021 TNF-alpha -170 6 -170a 0.005 TNF-alpha -120 6 -120a 0.011 Bouchard et al. Obese postmenopausal 14 SAT 6-month WLI / high responders: lost ≥3% DNASE1L2 CpG15 16 UHNhscpg0028273b 0.005 (2010)56 women BF DNASE1L2 CpG16 16 UHNhscpg0028273b 0.005 DNASE1L2 CpG17 16 UHNhscpg0028273b 0.005 OAT CpG21 10 UHNhscpg0022618b 0.003 Milagro et al. (2011)57 Overweight/obese men 12 PBMCs 8-week WLI/ high responders: lost ≥5% W ATP10A CpG5-10-16 15 23577342/23577440c <0.05 ATP10A CpG18 15 23577595 c <0.05 CD44 CpG14 11 35117620 c <0.05 CD44 CpG29 11 35117876 c <0.05 Cordero et al. (2011)58 Obese women 27 SAT 8-week WLI/ high responders: lost ≥5% W Leptin 7 -454a 0.017d TNF-alpha 6 -245a/ -239a 0.071d Milagro et al. (2012)59 Normal-weight, 60 White blood 4-month WLI CLOCK CpG1 4 -611a 0.010d overweight and obese cells PER2 CpG2-3 2 +125a 0.016d women PER2CpG25 2 +375a 0.016d Moleres et al. (2013)60 Overweight/obese 107 Blood 10-week intensive lifestyle intervention / AQP9 CpG1 15 56217683 c <0.05 adolescents high responders: lost > 1.1 BMI-SDS AQP9 CpG7 15 56217683 c <0.05 HIPK3 CpG1 11 33264921 c <0.05 HIPK3 CpG5 11 33264921 c <0.05 HIPK3 CpG7 11 33264921 c <0.05 HIPK3 total methylation 11 33264921 c <0.05 TNNI3 CpG1 19 60360424 c <0.05 TNNI3 CpG3 19 60360424 c <0.05 Pattern methylation after the weight loss intervention Milagro et al. (2011)57 Overweight/obese men 12 PBMCs 8-week WLI/ high responders: lost ≥5% W WT1 CpG2-3 11 32405898 c <0.05 WT1 CpG4 11 32405931 c <0.05 WT1 CpG7-8 11 32405983 c <0.05 WT1 CpG9-10 11 32406026 c <0.05 WT1 CpG13-32 11 32406078 c <0.05 WT1 CpG14 11 32406097 c <0.05 WT1 CpG15 11 32406116 c <0.05 WT1 CpG16-17 11 32406149 c <0.05 WT1 CpG21 11 32406214 c <0.05 Abbreviations: BF, body fat; BMI-SDS, body mass index-standard deviation score; Chr, chromosome; PBMCs, Peripheral blood mononuclear cells; SAT, Subcutaneous adipose tissue; W, Weight; WLI, Weight loss intervention. a Distance to the transcription start site. b Probe set. c Chromosome CpG position. d P value of Pearson’s correlation.

111

CHAPTER 2

Future perspectives of personalized weight-loss interventions based on nutrigenetic, epigenetic, and metagenomic data

Goni L.1,2, Cuervo M.1,2,3,4, Milagro F.I.1,2,3,4 ,Martínez J.A.1,2,3,4

1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Navarra, Spain 2 Centre for Nutrition Research, University of Navarra, Navarra, Spain 3 Navarra Institute for Health Research (IdiSNA), Navarra, Spain 4 Biomedical Research Centre Network in Physiopathology of Obesity and Nutrition (CIBERobn), Institute of Health Carlos III, Madrid, Spain

J Nutr. 2016 DOI 10.3945/jn.115.218354 Impact factor (2016): 4.145 16/81 Nutrition & Dietetics, Q1

Goni L, Cuervo M, Milagro FI, ,Martinez J.A. Future perspectives of personalized weight-loss interventions based on nutrigenetic, epigenetic, and metagenomic data. The Journal of Nutrition, 2016, 146(4):9055-9125. https://doi.org/10.3945/jn.115.218354

CHAPTER 3

Validation of a food groups frequency questionnaire based in an exchange system

Goni L.1,2, Aray M. 1,2, Martínez J.A.1,2,3,4, Cuervo M.1,2,3,4,

1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Navarra, Spain 2 Centre for Nutrition Research, University of Navarra, Navarra, Spain 3 Navarra Institute for Health Research (IdiSNA), Navarra, Spain 4 Biomedical Research Centre Network in Physiopathology of Obesity and Nutrition (CIBERobn), Institute of Health Carlos III, Madrid, Spain

Nutr Hosp. 2016 DOI 10.20960/nh.800 Impact factor (2016): 0.747 68/81 Nutrition & Dietetics, Q4

Goni L, Aray, M, Martinez JA, Cuervo M. Validation of food groups frequency questionnaire based in an exchange system. Nutrición hospitalaria, 2016, 33:1391- 1399. https://doi.org/10.20960/nh.800

ORIGINAL RESEARCH

CHAPTER 4

A genetic risk tool for obesity predisposition assessment and personalized nutrition implementation based on macronutrient intake

Goni L.1,2, Cuervo M.1,2,3,4, Milagro F.I.1,2,3,4, Martínez J.A.1,2,3,4

1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Navarra, Spain 2 Centre for Nutrition Research, University of Navarra, Navarra, Spain 3 Navarra Institute for Health Research (IdiSNA), Navarra, Spain 4 Biomedical Research Centre Network in Physiopathology of Obesity and Nutrition (CIBERobn), Institute of Health Carlos III, Madrid, Spain

Genes Nutr. 2015 DOI 10.1007/s12263-014-0445-z Impact factor (2015): 2.398 41/80 Nutrition & Dietetics, Q3 88/166 Genetics & Heredity, Q3

Results

163 Chapter 4

ABSTRACT

There is little evidence about genetic risk score (GRS)-diet interactions in order to provide personalized nutrition based on the genotype. The aim of the study was to assess the value of a GRS on obesity prediction and to further evaluate the interactions between the GRS and dietary intake on obesity. A total of 711 seekers of a Nutrigenetic Service were examined for anthropometric and body composition measurements and also for dietary habits and physical activity. Oral epithelial cells were collected for the identification of 16 SNPs (related with obesity or lipid metabolism) using DNA zip-coded beads. Genotypes were coded as 0, 1 or 2 according to the number of risk alleles and the GRS was calculated by adding risk alleles with such a criterion. After being adjusted for gender, age, physical activity and energy intake, the GRS demonstrated that individuals carrying > 7 risk alleles had in average 0.93 kg/m2 of BMI, 1.69% of body fat mass, 1.94 cm of waist circumference and 0.01 waist to height ratio more than the individuals with ≤7 risk alleles. Significant interactions for GRS and the consumption of energy, total protein, animal protein, vegetable protein, total fat, saturated fatty acids, polyunsaturated fatty acids, total carbohydrates, complex carbohydrates and fiber intake on adiposity traits were found after adjusted for confounders variables. The GRS confirmed that the high genetic risk group showed greater values of adiposity than the low risk group, and demonstrated that macronutrient intake modifies the GRS association with adiposity traits.

Keywords Genetic risk score; Obesity; Adiposity; Gene-macronutrient interaction

164 Results

1. INTRODUCTION

The prevalence of obesity is rising steadily not only in high-income countries but also in low- income countries. Indeed, it has been estimated that 1.12 billion adults will be obese by 2030 (Kelly et al. 2008). Consequently, the prevalence of obesity-associated metabolic diseases such as type 2 diabetes, cardiovascular disease or certain cancers, will also increase (Wang et al. 2011).

Although obesity is generally attributed to an imbalance between the energy intake and the energy expenditure, heritability studies indicate that genetic factors also play an important role in energy metabolism and the susceptibility to obesity (Abete et al. 2010; Min, Chiu, and Wang 2013). Indeed, genome wide association studies (GWAS) have identified a large number of single nucleotide polymorphisms (SNPs) associated with obesity and metabolic-related traits, which later have been widely replicated in different populations (Fall and Ingelsson 2014; Lind and Chiu 2013; Qi and Hu 2012; Global Lipids Genetics Consortium et al. 2013). Previous simulation studies have shown that the predictive value of a genetic variant could be improved by combining multiple loci simultaneously in a genetic risk score (GRS) model (Moonesinghe, Liu, and Khoury 2010). In fact, (Belsky et al. 2013) concluded that a 32 obesity locus GRS is a good predictor of body mass index (BMI) and obesity among Caucasians. Although it has been reported the discriminative and predictive power of GRSs of obesity SNPs, to the best of our knowledge, no study so far has specifically evaluated a multi-trait GRS of both, obesity and lipid metabolism SNPs (San-Cristobal et al. 2013).

Genetic factors could account for 25-70% of the body weight inter-individual variability (Razquin et al. 2011). Thus, genetic variants alone seem to be insufficient to explain the obesity heritability. In this context, a large number of investigations have suggested several interactions between certain genetic polymorphisms and modifiable environmental factors such as dietary intake or physical activity (Qi 2014; Roth et al. 2012). Nevertheless, there is scarce information regarding interactions between obesity GRSs and lifestyle factors. Interactions with physical activity (Li et al. 2010; Ahmad et al. 2013), television watching (Qi et al. 2012), meal frequencies (Jaaskelainen et al. 2013), omega-3 polyunsaturated fatty acids (PUFA) (Lemas et al. 2013), sugar-sweetened beverages (Qi et al. 2012) and fried food (Qi et al. 2014) have been described, although other report could not observe any strong interactions with macronutrients (Rukh et al. 2013).

165 Chapter 4

Thus, the two main aims of this study were to test the associations between a multi-trait GRS and body composition measurements; and to further evaluate dietary intake and obesity interactions depending on the genotype.

2. MATERIALS AND METHODS

2.1. Subjects

The study population included men and women of Caucasian ancestry who voluntarily attended a Nutrigenetic Service located in community pharmacies in seven regions of Spain (Barcelona, Zaragoza, La Coruña, Pontevedra, Madrid, Granada and Málaga). Genotype information of 718 individuals was available. Of these, 7 subjects were excluded with missing values for dietary intake, physical activity and/or anthropometric measurements. In total, 711 individuals were included in the analysis.

Seekers of the Nutrigenetic Service were specifically asked if they would be willing to take part anonymously in the research study. After ensuring that participants had understood the information, only those who provided written informed consent for participation were enrolled. The survey was in accordance with the principles of the Declaration of Helsinki and patient data were codified to guarantee anonymity accuracy (World Medical Association 2013). The Research Ethics Committee of the University of Navarra gave confirmation of fulfillment of the ethical standards and deontological criteria affecting this research (ref. 2710/2014).

2.2. Data collection

Anthropometric measurements were collected by trained nutritionists using a standardized protocol. Individuals were weighed with a digital scale (TANITA BF522W), wearing light clothes and no shoes. Height was measured using a portable stadiometer (Leicester TANITA) with subjects in barefoot. BMI was calculated dividing weight (kg) by the square of height (m). Body fat mass (BFM) was determined by bioelectrical impedance using the TANITA BF522W. Body fat distribution was evaluated by the measurement of waist and hip circumferences using a flexible and inextensible tape measure. Waist circumference was measured at the midway between the lower margin of the least rib and the top of iliac crest or according to the circumference at the level of the umbilicus if it was not possible to identify the least rib or the

166 Results iliac crest; and hip circumference as the widest circumference over the greater buttocks. Waist-to-hip ratio (waist/hip) and waist-to-height ratio (waist/height) were then calculated.

Physical activity was determined using a short 24 hours physical activity questionnaire (Panel on Macronutrients et al. 2005). Subjects were asked about the number of hours resting and practicing activities at work or at leisure time during a week day and a weekend day. Activities were divided in four groups according to intensity of effort: sedentary, low active, active and very active. Individual daily physical activity level was calculated multiplying the average time spent on each group of activities during the week and the weekend and the multiples of physical activity levels (Panel on Macronutrients et al. 2005).

Diet information was collected by a food frequency questionnaire in which basic foods were classified into 19 food groups: whole dairy products, half-fat dairy products and fat-free dairy products, eggs, fat meat and sausages, lean meat, white fish and selfish, blue fish, vegetables, fruits, nuts, legumes, olive oil, other fats and oils, refined grains, whole grains, pastries and confectionary industry, sugars, water and alcohol drinks. Each food group comprised several foods based on a food exchange list system (de la Iglesia et al. 2014). In other words, the foods on each group have about the same amount of calories, carbohydrate, protein and fat. For example, the group of whole dairy products (in average 67.6 kcal, 3.9 g carbohydrate, 4.2 g protein and 4.0 g fat) comprised 125 ml of milk, 50 g of milk powder, 125 g of different types of yogurt and 50 g of fresh cheese. Subjects were asked to report how often (daily, weekly, monthly or never), they had consumed a choice of each food group during the previous year. Total energy intake and macronutrient composition were calculated using Spanish food composition tables (Mataix et al. 2009; Moreiras et al. 2012).

2.3. Genetic risk score (GRS)

This research was based on a genetic tool develop by a pharmaceutical company (CINFA) which comprised 23 polymorphisms related with nutrition. Of the 23 SNPs a total of 16 SNPs previously associated with obesity and lipid metabolism on the basis of published reports were selected. Of these, rs9939609 and rs17782313, which represent the obesity susceptibility loci in the FTO and MC4R genes, respectively; have been identified by GWAS (Frayling et al. 2007; Willer et al. 2009; Loos et al. 2008). A recent meta-analysis has confirmed the association between the genetic variant rs1801282 (PPARG) and BMI (Galbete et al. 2013). The polymorphisms rs1801133 (MTHFR) and rs894160 (PLIN1) have been related with obesity phenotype in some association studies (Lewis et al. 2008; Soenen et al. 2009). Seven

167 Chapter 4 polymorphisms, rs1260326 (GCKR), rs662799 (APOA5), rs4939833 (LIPG), rs1800588 (LIPC), rs328 (LPL), rs12740374 (CELSR2) and rs7412 (APOE); have been reported to be associated with some lipid metabolism disturbances (hypertriglyceridemia, hypercholesterolemia, high low density lipoprotein -LDL- levels or low high density lipoprotein -HDL- levels) by GWAS (Willer et al. 2008; Kathiresan et al. 2008; Sandhu et al. 2008; Kettunen et al. 2012). The SNP rs429358 (APOE) has been associated with some lipid metabolism disorders by a meta-analysis of association studies (Bennet et al. 2007); and three SNPs, rs1799983 (NOS3) rs1800777 (CETP) and rs1800206 (PPARA), by association studies (Ferguson et al. 2010; Chrysohoou et al. 2004; Lu et al. 2008; Tai et al. 2002).

The GRS was calculated assuming that each SNP acts independently and contributes equally to the risk of obesity in an additive manner, as has been previously reported (He et al. 2010; Peterson et al. 2011). Genotypes were coded as 0, 1 or 2 according to the number of risk alleles for each variant. The GRS was computed by summing the risk alleles across the 16 SNPs for each individual.

2.4. Genotyping

Genomic DNA from oral epithelial cells (collected in ORAcollect DNA®, DNAGenotek) was extracted by QIAcube using QiAmp DNA Mini QIAcube Kit (Qiagen), following the manufacturer procedures. The polymerase chain reactions (PCRs) were carried out using the GeneAmp® PCR System 9700 thermal cycler according to standardized laboratory protocols. PCR products were analyzed using Luminex® 100/200TM System, which is based on the principles of xMAP® Technology. Briefly, this method uncompressed polystyrene microspheres internally dyed with various ratios of spectrally distinct fluorophores, which are detected by a flow cytometry-based instrument (Dunbar 2006).

2.5. Statistical analysis

Deviation from Hardy-Weinberg equilibrium was tested by χ2 test. Linear regression analyses (adjusted for age, sex, physical activity and total energy intake) were used to evaluate the association between the GRS or individual SNPs and body composition measurements (BMI, percentage of BFM, waist circumference, hip circumference, waist/hip and waist/height). Logistic regression analyses were applied to examine odds ratios (ORs) of GRS for obesity (BMI≥ 30 kg/m2; percentage of BFM >25% for males and >33% for females), and for risk of cardiovascular disease (waist circumference >94 cm for males and >80 cm for females;

168 Results waist/hip ≥0.90 for males and ≥0.85 for females; waist/height ≥0.53 for males and ≥0.51 for females), after they were adjusted for age, sex, physical activity and total energy intake as confounder variables (Rubio et al. 2007; Sociedad Española para el Estudio de la Obesidad - SEEDO- 2000; World Health Organization 2008). The discriminative power of the GRS was tested by the area under the receiver operating characteristic curves (ROC AUCs) (Dorfman and Alf Jr 1969; Cleves 1999). Interactions between the GRS or SNPs and diet intake on percentage of BFM and obesity risk (according to the percentage of BFM) were examined with the likelihood ratio test (Sánchez-Villegas and Martínez-Gonzalez 2006). Product terms between the GRS or SNPs and diet intake were calculated with the nutrients dichotomized at the median and as continuous variables. All statistical analyses were performed using STATA/SE version 12.0 (StataCorp, College Station, TX, USA). A p value of p<0.05 was considered as statistically significant.

3. RESULTS

The phenotypical characteristics of the individuals included in the study were categorized by gender (Table 1). As expected, males showed higher height, weight and waist circumference than females. In contrast, females had greater percentage of BFM. Energy intake and physical activity were higher in males than in females. Of the 711 subjects 41.6% were obese, and no statistical differences were observed depending on gender. About 40.9% of the individuals self-declared that they suffered one or more metabolic disorders: 22.2% hypertension, 3.2% type 2 diabetes, 28.0% different lipid metabolism impairments and 3.1% cardiovascular disease.

The minor allele frequencies (MAF) and Hardy Weinberg Equilibrium for each SNP are listed (Table 2). MAF ranged from 0.02 to 0.45 in the population. The distributions of the all polymorphisms alleles were in Hardy Weinberg Equilibrium (p>0.05), except rs1800777 in CETP (p=0.0052). Moreover, risk alleles for each SNP and alleles frequencies in Hap-Map CEU population have been reported (Supplementary Material Table 1).

The association of each SNP of the GRS with obesity-related traits including BMI, percentage of BFM, waist circumference, waist/hip and waist/height (after adjusted for age, gender, physical activity and energy intake) was analyzed (data not shown). Individuals with the genotype AA of the rs9939609 genetic variant in the FTO gene had greater BMI, percentage of BFM, waist circumference and waist/height. The C allele carriers of the genetic variant rs429358 (APOE) showed significantly higher BMI and a tendency was observed for the percentage of BFM,

169 Chapter 4 waist circumference and hip circumference. None of the other SNPs showed significant association with any other quantitative traits in the case of the total population. After adjusted for confounder variables, only rs9939609 (FTO) was significantly associated with obesity risk in the current study, while rs429358 (APOE) showed a trend toward significance (Supplementary Material Table 1).

The GRS, calculated as the number of risk alleles carried by each subject, was normally distributed. The average number of risk alleles per person was 8.09 (SD=2.02), which ranged from 2 to 16. The sample was stratified, by the median, into a “low genetic risk group”, those with a GRS ≤7 risk alleles (n=289), and a “high genetic risk group”, those with a GRS >7 risk alleles (n=422).

Results from linear regression and logistic regression analyses revealed that the group with a GRS of >7 risk alleles, had a 0.93 kg/m2 greater BMI, a 1.69 greater percentage of BFM, a 1.94 cm larger waist circumference and a greater 0.01 waist/height compared with the group with a GRS of ≤7 risk alleles, after they were adjusted for age, gender, physical activity and energy intake (Table 3). It should be highlighted that when the GRS was considered as a continuous variable, the associations between BMI, waist circumference, waist/hip and waist/height ratios, and GRS were no longer statistically significant. However, a significant association was observed with the percentage of BFM when considered the GRS as a continuous variable (B coefficient 0.373, 95%CI 0.10-0.64, p 0.007), after adjusted for the previously mentioned confounder variables. Individuals in the high genetic risk group, after confounder variables adjustment (age, gender, physical activity and energy intake), were at 42 and 72% greater risk of obesity according to BMI and percentage of BFM, respectively. The risk of cardiovascular disease, taking into account waist circumference, was 57% higher in the high genetic risk group than in the low genetic risk group.

Moreover, the ROC curves for the prediction of obesity according to BMI and the percentage of BFM were calculated. The AUC estimates were low. The GRS predicted obesity with a maximum discriminating ability when the percentage of BFM was taken into account (AUC 0.55; 95% CI 0.51-0.60). The AUC was 0.53 (95% CI 0.48-0.57) when the BMI was considered.

Significant SNPs interactions (FTO, PPARG, APOE) with some macronutrient intake that modified the association with BFM were detected (data not shown). After correction for multiple testing, interactions between FTO and total fat (p interaction=0.016) and MUFA intake (p interaction=0.012), PPARG and animal protein (p interaction=0.011) and diet

170 Results cholesterol intake, (p interaction=0.028) and APOE and complex carbohydrates intake (p interaction=0.027) on BFM remained significant.

After data were adjusted for gender, age, physical activity and energy intake, several macronutrients modified the effect of the GRS on percentage of BFM and risk of obesity (Table 4, Supplementary Material Fig. 1-4). A higher energy intake was associated with an increase of 0.39 (SE 0.04) percentage of BFM in the high genetic risk group, while the effect was much less in the low genetic risk group. Similar trends for interaction were found for protein (p interaction=0.003), and fat intakes (p interaction=0.029). Interestingly, when stratifying by animal and vegetable protein, the vegetable intake seemed to improve the percentage of the BFM (p interaction=0.003) in the ≤7 risk alleles group, whereas the animal protein intake was associated with an increase in the percentage of BFM in the >7 risk alleles group (p interaction=0.032). Moreover, the higher the increase in AGS (p interaction=0.015), carbohydrates (p interaction=0.008) and complex carbohydrates (p interaction=0.030), the higher the percentage of BFM in the high genetic risk group. In contrast, the higher intake of AGP (p interaction=0.002) and fiber (p interaction=0.039), the lower the percentage of BFM in the low genetic risk group. The same tendency was observed in those analyses taking into account the risk of obesity as the percentage of BFM (Table 4). Similar results were evidenced regarding interactions between dietary intakes and GRS, by dichotomizing macronutrients according to the mean population intakes and evaluating their effects on the percentage of BFM as continuous and categorical variable (adjusted for age, gender, physical activity and energy intake) (data not shown).

4. DISCUSSION

In this cross-sectional study encompassing 711 users of a Nutrigenetic Service, it was investigated the predictive value of 16 obesity and lipid metabolism risk polymorphisms in the adiposity outcome and the possible modifying effect of the dietary intake. Our study shows that after combining information of 16 polymorphisms, those people who were classified as high genetic risk group presented higher values in obesity related traits and a higher risk of obesity than the alternative group. In addition, the consumption of some specific macronutrients modified the association between the GRS and BFM and obesity risk. Interestingly, this is apparently the first study analyzing whether macronutrient intake can modulate the association between a multi-trait GRS and obesity.

171 Chapter 4

When looking at single SNPs results, in agreement with previous investigations, the current study reported a significant association between rs9939609 (FTO) and BMI, BFM and measures of central obesity (waist circumference and waist/height ratio) (Fall and Ingelsson 2014). Recently, it has been published that FTO is functionally connected with the regulation of IRX3 expression, which is involved in body composition (Smemo et al. 2014). Although the relationship between FTO and obesity-related traits has been confirmed in several populations, in Spanish individuals there is a little controversy (Fall and Ingelsson 2014). One study investigated the contribution of FTO, among other several polymorphisms, to obesity in two Spanish cohorts (Martinez-Garcia et al. 2013). Whereas FTO was associated with BMI in the Pizarra cohort, this result was not found in the Hortega cohort (Martinez-Garcia et al. 2013). In addition to the well-described rs9939609 (FTO), the polymorphism rs429358 located in APOE showed a statistical significant association with BMI in the present study. APOE gene encodes for a major structural apolipoprotein of several lipoprotein classes (chylomicrons, very low density lipoprotein -VLDL- and HDL), which plays diverse roles in lipoprotein metabolism. Polymorphisms in APOE have been previously related with lipid metabolism disorders, cardiovascular disease, Alzheimer’s disease and obesity phenotype (Sima, Iordan, and Stancu 2007; Kypreos et al. 2009; Ridge, Ebbert, and Kauwe 2013). Although a large number of studies have observed the association between APOE genotypes and obesity, up to date no study has found any statistically significant association between rs429358 and obesity phenotype (Clark et al. 2009). A number of studies in vitro and in vivo have explained the link between APOE and obesity, due to the contribution of APOE in adipogenesis (Kypreos et al. 2009).

A weighted GRS, computed by weighting each variant by their effect size, can be used. However, we selected a simple count method because weighted models may have only limited effects, as has been reported by some previous studies (Belsky et al. 2013; Cheung et al. 2010). Moreover, the effect sized on obesity for some of the selected SNPs is not established. The GRS design was based on the premise that some of the loci reliably associated with obesity- related traits, such as lipid metabolism disorders, might increase the power to discriminate between individuals with and without obesity. Although the polymorphism rs1800777 (CETP) was not in Hardy-Weinberg Equilibrium it was included in the GRS because it was associated with the percentage of BFM in males and females.

The GRS was positively associated with several measures of obesity in our study population. In this regard, it should be noted that from a clinical point of view, the categorization of a genetic score into a dichotomous variable could be more useful for clinical decision-making (Horne et al. 2005). The analyses demonstrated that the GRS as a continuous variable predicts <1% of the

172 Results

BMI variation (0.16%) in concordance with most of previous studies (Li et al. 2010; Takeuchi et al. 2011; Peterson et al. 2011; Lemas et al. 2013). Nevertheless, our GRS explained 1.59%. Although BMI has been widely used as a surrogate measure of adiposity, it does not distinguish between BFM and lean mass. So, the accuracy of BMI in assessing the BFM remains debated and the measurement of BFM is a more specific measure of adiposity than BMI, which can explain our findings (Bergman et al. 2011).

Additionally, ROC curves and the corresponding AUC estimates indicated statistical discriminative ability to predict obesity taking into account both, BMI and the percentage of BFM. The discriminatory ability of our GRS was similar when compared to other obesity GRS, which ranged between 0.575 and 0.697 (Renstrom et al. 2009; Cheung et al. 2010; Li et al. 2010; Peterson et al. 2011; Belsky et al. 2013).

Our data show that high energy intake was associated with higher percentage of BFM in the population overall, but that in particular individuals of the high genetic risk group the effect on adiposity of a high energy intake was stronger. So, we decided to adjust for energy intake the rest of the analyses interactions between macronutrients and GRS.

Currently, the beneficial effect of the consumption of proteins is debated, mainly in relation with the protein source (Lopez-Legarrea et al. 2014). Higher total protein and animal protein intake were significantly associated with higher BFM in both genetic risk groups, but this association was significantly stronger among individuals of the high genetic risk group. Meanwhile, vegetable protein intake appears to have a protective effect among individuals of the low genetic risk group. Previously Rukh et al. found that protein intake modulates the association between a 13 polymorphisms GRS and obesity and fat mass in Swedish women (Rukh et al. 2013). Also a gene-diet interaction between PPARG and total protein and animal protein intake was observed in BFM. In this sense, other investigations have previously reported the PPARG interaction with total fat, MUFA and carbohydrate intake in obesity- related traits (Robitaille et al. 2003; Dedoussis et al. 2011; Garaulet et al. 2011; Galbete et al. 2013).

Although the association between total fat and SFA intake and obesity is well-studied, this is the first investigation confirming the combined effect of these nutrients and GRS on adiposity (Cascio, Schiera, and Di Liegro 2012). In this context, Qi et al. (2014) reported that among individuals with a higher GRS, the association between fried food consumption and BMI was stronger than among individuals with a lower GRS. Our study also suggests that PUFA intake could modify the genetic association with body composition. This is in accordance with

173 Chapter 4 findings from a previous study which detected a significant interaction between a 12 obesity SNPs’ GRS and omega-3 PUFA and BMI (Lemas et al. 2013). In addition to the interaction with GRS, several nominal interactions with FTO and dietary fat were found; thus FTO interacted with total fat and MUFA to modify its effect on BFM. Earlier studies have also reported FTO- diet interactions on obesity related traits (Sonestedt et al. 2009; Ahmad et al. 2011; Lappalainen et al. 2012; Phillips et al. 2012).

Another novel finding of our study is that total carbohydrates and complex carbohydrates interacted with the GRS to modify the effect on BFM. There are no previous reports regarding these interactions, but Qi et al. (2012b) found an interaction between sugar-sweetened beverages and a GRS of 32 obesity polymorphisms in 3 US cohort studies. Finally, an interaction between the polymorphism rs7412 (APOE) and complex carbohydrates was found.

The major strength of the present study is the use of a population who attended voluntarily a Nutrigenetic Service. As far as we know, this is the first study that analyses a real population of a nutritional service based on the genetic makeup. Moreover, we used a bioimpedance technique to evaluate body composition instead of BMI or other anthropometric measurements, as many studies do. Nevertheless, some limitations need to be acknowledged. The users included in the study were middle aged and of European ancestry, so it is unknown whether our results can be generalized to other demographic or ethnic groups. It is possible that the sample size might be rather relatively small; however consistent GRS-environment interactions were found. Furthermore, the 16 selected SNPs may be only markers of functional variations and that future GWAS data will contribute to a better understanding of the genetic background of this population.

In conclusion, this study provides interesting novel information on obesity and lipid metabolism-related genes, their additive effects on the risk of obesity, and the modulation by dietary habits. The investigation of GRS-diet interactions may facilitate the selection of more individualized effective nutritional therapy, following personalized approaches based on the genotype.

Acknowledgements

The authors thank all the users of the Nutrigenetic Service who voluntarily offer her data to the study. The authors are grateful to the 5 nutritionists for data collection and to Amaia Ibañez for excellent technical assistance. The predoctoral research grant to Leticia Goni from the Asociación de Amigos Universidad de Navarra is gratefully acknowledged. The authors also

174 Results wish to thank the Linea Especial (University of Navarra; LE/97) for financial support and also CIBERobn/RETICS schedules (Instituto Carlos III) for assistance in this study. The support from CINFA concerning the genetic tools and general logistic is also gratefully acknowledgement.

Conflict of interest

The authors declare no financial conflict of interest concerning this research. However, there are some results of this work that could be incorporated to the future development of new nutrigenetic tests.

Ethical standard

All procedures were in accordance with the Helsinki declaration 1975, as revised in 2010. Informed consent was obtained from all patients before for being included in the study.

175 Chapter 4

REFERENCES

Abete I, Astrup A, Martinez JA, Thorsdottir I, Zulet MA (2010) Obesity and the metabolic syndrome: Role of different dietary macronutrient distribution patterns and specific nutritional components on weight loss and maintenance. Nutr Rev 68:214-231. Ahmad S, Rukh G, Varga TV et al (2013) Gene x physical activity interactions in obesity: Combined analysis of 111,421 individuals of European ancestry. PLoS Genet 9:e1003607. Ahmad T, Lee IM, Pare G, Chasman DI, Rose L, Ridker PM, Mora S (2011) Lifestyle interaction with fat mass and obesity-associated (FTO) genotype and risk of obesity in apparently healthy U.S. women. Diabetes Care 34:675-680. Belsky DW, Moffitt TE, Sugden K, Williams B, Houts R, McCarthy J, Caspi A (2013) Development and evaluation of a genetic risk score for obesity. Biodemography Soc Biol 59:85-100. Bennet AM, Di Angelantonio E, Ye Z et al (2007) Association of Apolipoprotein E genotypes with lipid levels and coronary risk. JAMA 298:1300-1311. Bergman RN, Stefanovski D, Buchanan TA, Sumner AE, Reynolds JC, Sebring NG, Xiang AH, Watanabe RM (2011) A better index of body adiposity. Obesity (Silver Spring) 19:1083-1089. Cascio G, Schiera G, Di Liegro I (2012) Dietary fatty acids in metabolic syndrome, diabetes and cardiovascular diseases. Curr Diabetes Rev 8:2-17. Cheung CY, Tso AW, Cheung BM, Xu A, Ong KL, Fong CH, Wat NM, Janus ED, Sham PC, Lam KS (2010) Obesity susceptibility genetic variants identified from recent genome-wide association studies: Implications in a Chinese population. J Clin Endocrinol Metab 95:1395-1403. Chrysohoou C, Panagiotakos DB, Pitsavos C, Antoniades C, Skoumas J, Brown M, Stefanadis C (2004) Evidence for association between endothelial nitric oxide synthase gene polymorphism (G894T) and inflammatory markers: The ATTICA study. Am Heart J 148:733-738. Clark D, Skrobot OA, Adebiyi I, Susce MT, de Leon J, Blakemore AF, Arranz. MJ (2009) Apolipoprotein-E gene variants associated with cardiovascular risk factors in antipsychotic recipients. Eur Psychiatry 24:456-463. Cleves M (1999) Receiver Operating Characteristic (ROC) analysis. Stata technical bulletin STB-52. de la Iglesia R, Lopez-Legarrea P, Abete I, Bondia-Pons I, Navas-Carretero S, Forga L, Martinez JA, Zulet MA (2014) A new dietary strategy for long-term treatment of the metabolic syndrome is compared with the American Heart Association (AHA) Guidelines: The MEtabolic Syndrome REduction in NAvarra (RESMENA) project. Br J Nut 111:643- 652. Dedoussis GV, Manios Y, Kourlaba G et al. (2011) An age-dependent diet-modified effect of the PPARgamma Pro12Ala polymorphism in children. Metabolism 60:467-473. Dorfman DD, Alf Jr E (1969) Maximum-likelihood estimation of parameters of signal-detection theory and determination of confidence intervals-rating method data. J Math Psychol 6:487-496. Dunbar SA (2006). Applications of Luminex xMAP technology for rapid, high-throughput multiplexed nucleic acid detection. Clin Chim Acta 363:71-82. Fall T, Ingelsson E (2014) Genome-wide association studies of obesity and metabolic syndrome. Mol Cell Endocrinol 382:740-757. Ferguson JF, Phillips CM, McMonagle J et al (2010) NOS3 gene polymorphisms are associated with risk markers of cardiovascular disease, and interact with omega-3 polyunsaturated fatty acids. Atherosclerosis 211:539-544. Frayling TM, Timpson NJ, Weedon MN et al (2007) A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science 316:889-894.

176 Results

Galbete C, Toledo E, Martinez-Gonzalez MA, Martinez JA, Guillen-Grima F, Marti A (2013) Pro12Ala variant of the PPARG2 gene increases body mass index: An updated meta-analysis encompassing 49,092 subjects. Obesity (Silver Spring) 21:1486-1495. Galbete C, Toledo J, Martinez-Gonzalez MA, Martinez JA, Guillen-Grima F, Marti A (2013) Lifestyle factors modify obesity risk linked to PPARG2 and FTO variants in an elderly population: A cross-sectional analysis in the SUN project. Genes Nutr 8:61-67. Garaulet M, Smith CE, Hernandez-Gonzalez T, Lee YC, Ordovas JM (2011) PPARgamma Pro12Ala interacts with fat Intake for obesity and weight loss in a behavioural treatment based on the Mediterranean diet. Mol Nutr Food Res 55:1771-1779. Global Lipids Genetics Consortium, Willer CJ, Schmidt EM et al (2013) Discovery and refinement of loci associated with lipid levels. Nat Genet 45:1274-1283. He M, Cornelis MC, Franks PW, Zhang C, Hu FB, Qi L (2010) Obesity genotype score and cardiovascular risk in women with type 2 diabetes mellitus. Arterioscler Thromb Vasc Biol 30:327-332. Horne BD, Anderson JL, CarlquistJF, Muhlestein JB, Renlund DG, Bair TL, Pearson RR, Camp NJ (2005) Generating genetic risk scores from intermediate phenotypes for use in association studies of clinically significant endpoints. Ann Hum Genet 69:176-186. Jaaskelainen A, Schwab U, Kolehmainen M, Kaakinen M, Savolainen MJ, Froguel P, Cauchi S, Jarvelin MR, Laitinen J (2013) Meal frequencies modify the effect of common genetic variants on body mass index in adolescents of the Northern Finland Birth Cohort 1986. PloS One 8:e73802. Kathiresan S, Melander O, Guiducci C et al (2008) Six new loci associated with blood low-density lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in humans. Nat Genet 40:189-197. Kelly T, Yang W, Chen CS, Reynolds K, He J (2008) Global burden of obesity in 2005 and projections to 2030. Int J Obes 32:1431-1437. Kettunen J, Tukiainen T, Sarin AP et al (2012) Genome-wide association study identifies multiple loci influencing human serum metabolite levels. Nat Genet 44:269-276. Kypreos KE, Karagiannides I, Fotiadou EH, Karavia EA, Brinkmeier MS, Giakoumi SM, Tsompanidi EM (2009) Mechanisms of obesity and related pathologies: role of Apolipoprotein E in the development of obesity. FEBS J 276:5720-5728. Lappalainen T, Lindstrom J, Paananen J, Eriksson JG, Karhunen L, Tuomilehto J, Uusitupa M (2012) Association of the fat mass and obesity-associated (FTO) gene variant (rs9939609) with dietary intake in the Finnish Diabetes Prevention Study. Br J Nut 108:1859-1865. Lemas DJ, Klimentidis YC, Wiener HH, O'Brien DM, Hopkins SE, Allison DB, Fernandez JR, Tiwari HK, Boyer BB (2013) Obesity polymorphisms identified in genome-wide association studies interact with n-3 polyunsaturated fatty acid intake and modify the genetic association with adiposity phenotypes in Yup'Ik people. Genes Nutr 8:495-505. Lewis SJ, Lawlor DA, Nordestgaard BG, Tybjaerg-Hansen A, Ebrahim S, Zacho J, Ness A, Leary S, Smith GD (2008) The methylenetetrahydrofolate reductase C677T genotype and the risk of obesity in three large population-based cohorts. Eur J Endocrinol 159:35-40. Li S, Zhao JH, Luan J, Ekelund U,. Luben RN, Khaw KT, Wareham NJ, Loos RJ (2010) Physical activity attenuates the genetic predisposition to obesity in 20,000 men and women from EPIC-Norfolk prospective population study. PLoS Med 7:10.1371.

177 Chapter 4

Li S, Zhao JH, Luan J, Luben RN, Rodwell SA, Khaw KT, Ong KK,. Wareham NJ, Loos RJ (2010) Cumulative effects and predictive value of common obesity-susceptibility variants identified by genome-wide association studies. Am J Clin Nutr 91:184-190. Lind JM, Chiu CL (2013) Genetic discoveries in hypertension: steps on the road to therapeutic translation. Heart 99:1645-1651. Loos RJ, Lindgren CM, Li S et al (2008) Common variants near MC4R are associated with fat mass, weight and risk of obesity. Nat Genet 40:768-775. Lopez-Legarrea P, de la Iglesia R, Abete I, Navas-Carretero S, Martinez JA, Zulet MA (2014) The protein type within a hypocaloric diet affects obesity-related inflammation: The RESMENA project. Nutrition 30:424-429. Lu Y, Dolle ME, Imholz S, van 't Slot R, Verschuren WM, Wijmenga C, Feskens EJ, Boer JM (2008) Multiple genetic variants along candidate pathways influence plasma high-density lipoprotein cholesterol concentrations. J Lipid Res 49:2582-2589. Martinez-Garcia F, Mansego ML, Rojo-Martinez G, et al (2013) Impact of obesity-related genes in Spanish population. BMC Genet 14:111. Mataix J, García L, Mañas M, Martínez E, Llopis J (2009) Tablas de composición de alimentos, 5th edn. Universidad de Granada, Granada. Min J, Chiu DT, Wang Y (2013) Variation in the heritability of body mass index based on diverse twin studies: A systematic review. Obes Rev 14:871-882. Moonesinghe R, Liu T, Khoury MJ (2010) Evaluation of the discriminative accuracy of genomic profiling in the prediction of common complex diseases. Eur J Hum Genet 18:485-489. Moreira O, Carbajal A, Cabrera L, Cuadrado C (2012) Tablas de composición de alimentos, 15th edn. Pirámide, Madrid. Panel on Macronutrients, Panel on the Definition of Dietary Fiber, Subcommittee on Upper Reference Levels of Nutrients, Subcommittee on Interpretation and Uses of Dietary Reference Intakes, and the Standing Committee on the Scientific Evaluation of Dietary Reference Intakes (2005) Dietary reference intakes for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein, and amino acids (Macronutrients). The National Academies Press, Washington, DC. Peterson RE, Maes HH, Holmans P, Sanders AR, Levinson DF, Shi J, Kendler KS, Gejman PV, Webb BT (2011) Genetic risk sum score comprised of common polygenic variation is associated with body mass index. Hum Genet 129:221- 230. Phillips CM, Kesse-Guyot E, McManus R, Hercberg S, Lairon D, Planells R, Roche HM (2012) High dietary saturated fat intake accentuates obesity risk associated with the fat mass and obesity-associated gene in adults. J Nutr 142:824-831. Qi L (2014) Gene-diet interaction and weight loss. Curr Opin Lipidol 25:27-34. Qi Q, Chu AY, Kang JH et al (2014) Fried food consumption, genetic risk, and body mass index: gene-diet interaction analysis in three US cohort studies. BMJ 348:g1610. Qi Q, Chu AY, Kang JH et al (2012) Sugar-sweetened beverages and genetic risk of obesity. N Engl J Med 367:1387- 1396. Qi Q, Hu FB (2012) Genetics of type 2 diabetes in European populations. Journal of Diabetes 4:203-212. Qi Q, Li Y, Chomistek AK, Kang JH, Curhan GC, Pasquale LR, Willett WC, Rimm EB, Hu FB, Qi L (2012) Television watching, leisure time physical activity, and the genetic predisposition in relation to body mass index in women and men. Circulation 126:1821-1827.

178 Results

Razquin C, Marti A, Martinez JA (2011) Evidences on three relevant obesogenes: MC4R, FTO and PPARgamma. Approaches for personalized nutrition. Mol Nutr Food Res 55:136-149. Renstrom F, Payne F, Nordstrom A, Brito EC, Rolandsson O, Hallmans G, Barroso I, Nordstrom P, Franks PW, GIANT Consortium (2009) Replication and extension of genome-wide association study results for obesity in 4923 adults from Northern Sweden. Hum Mol Genet 18:1489-1496. Ridge PG, Ebbert MT, Kauwe JS (2013) Genetics of Alzheimer's disease. Biomed Res Int 2013:254954. Robitaille J, Despres JP, Perusse L, Vohl MC (2003) The PPAR-gamma P12A Polymorphism modulates the relationship between dietary fat intake and components of the metabolic syndrome: results from the Quebec Family Study. Clin Genet 63:109-116. Roth SM, Rankinen T, Hagberg JM, Loos RJ, Perusse L, Sarzynski MA, Wolfarth B, Bouchard C (2012) Advances in exercise, fitness, and performance genomics in 2011. Med Sci Sports Exerc 44:809-817. Rubio MA, Salas-Salvadó J, Barbany M et al (2007) Consenso SEEDO 2007 para la evaluación del sobrepeso y la obesidad y el establecimiento de criterios de intervención terapéutica. Rev Esp Obes:7-48. Rukh G, Sonestedt E, Melander O, Hedblad B, Wirfalt E, Ericson U, Orho-Melander M (2013) Genetic susceptibility to obesity and diet intakes: association and interaction analyses in the Malmo Diet and Cancer Study. Genes Nutr 8:535-547. Sánchez-Villegas A, Martínez-Gonzalez MA (2006) Aspectos avanzados de regresión múltiple. In: Martínez-González MA, Sánchez-Villegas A, Faulin J (ed) Bioestadística Amigable, 2nd edn. Diaz de Santos, España, pp:761-765. San-Cristobal R, Milagro FI, Martinez JA (2013) Future challenges and present ethical considerations in the use of personalized nutrition based on genetic advice. J Acad Nutr Diet 113:1447-1454. Sandhu MS, Waterworth DM, Debenham SL et al (2008) LDL-cholesterol concentrations: A genome-wide association study. Lancet 371:483-491. Sima A, Iordan A, Stancu C (2007) Apolipoprotein E polymorphism-a risk factor for metabolic syndrome. Clin Chem Lab Med 45:1149-1153. Smemo S, Tena JJ, Kim KH et al (2014) Obesity-associated variants within FTO form long-range functional connections with IRX3. Nature 507:371-375. Sociedad Española para el Estudio de la Obesidad (SEEDO) (2000) Consenso SEEDO'2000 para la evaluación del sobrepeso y obesidad y el establecimiento de criterios de intervención quirúrgica. Med Clin (Barc) 115:587-597. Soenen S, Mariman EC, Vogels N, Bouwman FG, den Hoed M, Brown L, Westerterp-Plantenga MS (2009) Relationship between Perilipin gene polymorphisms and body weight and body composition during weight loss and weight maintenance. Physiol Behav 96:723-728. Sonestedt E, Roos C, Gullberg B, Ericson U, Wirfalt E, Orho-Melander M (2009) Fat and carbohydrate intake modify the association between genetic variation in the FTO genotype and obesity. Am J Clin Nutr 90:1418-1425. Tai ES, Demissie S, Cupples LA, Corella D, Wilson PW, Schaefer WJ, Ordovas JM (2002) Association between the PPARA L162V polymorphism and plasma lipid levels: The Framingham Offspring Study. Arterioscl Thromb Vasc Biol 22:805-810. Takeuchi F, Yamamoto K, Katsuya T et al (2011) Association of genetic variants for susceptibility to obesity with type 2 diabetes in Japanese individuals. Diabetologia 54:1350-1359. Wang YC, McPherson K, Marsh T, Gortmaker SL, Brown M (2011) Health and economic burden of the projected obesity trends in the USA and the UK. Lancet 378:815-825. Willer CJ, Sanna S, Jackson AU et al (2008) Newly identified loci that influence lipid concentrations and risk of coronary artery disease. Nat Genet 40:161-169.

179 Chapter 4

Willer CJ, Speliotes EK, Loos RJ, Li S et al (2009) Six new loci associated with body mass index highlight a neuronal influence on body weight regulation. Nat Genet 41:25-34. World Health Organization (2008) Waist circumference and waist–hip ratio: Report of a WHO expert consultation. Geneva. World Medical Association (2013) World Medical Association Declaration of Helsinki: Ethical pinciples for medical research involving human subjects. JAMA 31:2191-2194.

180 Results

Table 1 Anthropometrical and nutritional baseline characteristics stratified and statistically compared by gender

Variables All (n=711) Males (n=159) Females (n=552) p-value Age (years) 50.1 ± 13.4 47.9 ± 13.6 50.7 ± 13.2 0.016 Height (cm) 163.0 ± 8.8 173.7 ± 7.5 159.9 ± 6.4 <0.001 Weight (kg) 78.2 ± 16.7 90.0 ± 16.4 74.8 ± 15.2 <0.001 BMI (kg/m2) 29.4 ± 5.8 29.8 ± 5.0 29.3 ± 6.0 0.339 Obesity (BMI ≥30kg/m2) (%) 296 (41.6) 68 (42.8) 228 (41.3) 0.742 a BFM (%) 34.6 ± 10.1 24.4 ± 9.7 37.5 ± 8.2 <0.001 Waist circumference (cm) 96.4 ± 15.2 104.3 ± 13.9 94.2 ± 14.7 <0.001 Waist/ hip 0.88 ± 0.09 0.96 ± 0.09 0.85 ± 0.08 <0.001 Waist/ height 0.59 ± 0.09 0.60 ± 0.08 0.59 ± 0.10 0.175 Energy intake (kcal) 2,151 ± 431 2,477 ± 464 2,057 ± 372 <0.001 Physical activity 1.23 ± 0.03 1.24 ± 0.03 1.23 ± 0.03 0.003 BMI, Body mass index; BFM, Body fat mass a Chi squared p value

181 Chapter 4

Table 2 Genotype, Minor Allele Frequency (MAF) and Hardy-Weinberg equilibrium calculations of the 16 SNPs included in the GRS

Major/minor Major allele Heterozygote Minor allele Gene SNP a MAF HWE p-value allele homozygote (%) (%) homozygote (%) FTO rs9939609 T/A 218 (30.6) 351 (49.4) 142 (20.0) 0.45 0.973 MC4R rs17782313 T/C 442 (62.2) 237 (33.3) 32 (4.5) 0.21 0.974 MTHFR rs1801133 C/T 257 (36.2) 340 (47.8) 114 (16.0) 0.40 0.930 PPARA rs1800206 C/G 591 (83.1) 116 (16.3) 3 (0.5) 0.09 0.507

PPARG rs1801282 C/G 594 (82.5) 110 (15.5) 7 (1.0) 0.09 0.453

APOA5 rs662799 T/C 620 (87.2) 88 (12.4) 3 (0.4) 0.07 0.948 APOE rs429358 T/C 574 (80.7) 129 (18.1) 8 (1.1) 0.10 0.804 APOE rs7412 C/T 624 (87.8) 84 (11.8) 3 (0.4) 0.06 0.923 LIPC rs1800588 C/T 406 (57.1) 250 (36.3) 47 (6.6) 0.24 0.489 PLIN1 rs894160 G/A 378 (53.2) 282 (39.7) 51 (7.2) 0.27 0.872

NOS3 rs1799983 G/T 286 (40.2) 326 (45.8) 99 (13.9) 0.37 0.692

GCKR rs1260326 C/T 208 (29.2) 367 (50.9) 141 (19.8) 0.45 0.465

LPL rs328 C/G 510 (71.7) 189 (26.6) 12 (1.7) 0.15 0.244 CELSR2 rs12740374 G/T 440 (61.9) 241 (33.9) 30 (4.2) 0.21 0.676 CETP rs1800777 G/A 681 (95.8) 28 (3.9) 2 (0.3) 0.02 0.005 LIPG rs4939883 C/T 508 (71.4) 179 (25.2) 24 (3.4) 0.16 0.100

SNP, Single nucleotide polymorphism; MAF, Minor allele frequency; HWE p-value, Hardy-Weinberg equilibrium p-value a According to Hap-Map CEU for European population

182 Results

Table 3 Linear regression coefficients and logistic regression coefficients for the association between GRS (dichotomized by the median) and several anthropometric variables

Linear regression coefficients Logistic regression coefficients Model 1 Model 2 Model 1 Model 2 B (95% CI) p-value B (95% CI) p-value OR (95% CI) p-value OR (95% CI) p-value BMI (kg/m2) ≤7 risk alleles 0 (ref.) 0 (ref.) 1 (ref.) 1 (ref.) >7 risk alleles 1.02 (0.17-1.87) 0.019 0.93 (0.17-1.68) 0.016 1.41 (1.04-1.93) 0.031 1.42 (1.02-1.99) 0.038 Percentage of BFM ≤7 risk alleles 0 (ref.) 0 (ref.) 1 (ref.) 1 (ref.) >7 risk alleles 1.81 (0.59-3.03) 0.004 1.69 (0.58-2.80) 0.003 1.72 (1.22-2.42) 0.002 1.72 (1.19-2.48) 0.004 Waist circumference (cm) ≤7 risk alleles 0 (ref.) 0 (ref.) 1 (ref.) 1 (ref.) >7 risk alleles 2.14 (0.08-4.21) 0.042 1.94 (0.12-3.75) 0.036 1.54 (1.04-2.29) 0.032 1.57 (1.02-2.40) 0.039 Waist to hip ratio ≤7 risk alleles 0 (ref.) 0 (ref.) 1 (ref.) 1 (ref.) >7 risk alleles 0.00 (-0.01-0.002) 0.445 0.00 (-0.01-0.01) 0.480 1.15 (0.82-1.61) 0.401 1.14 (0.81-1.61) 0.454 Waist to height ratio ≤7 risk alleles 0 (ref.) 0 (ref.) 1 (ref.) 1 (ref.) >7 risk alleles 0.01 (0.00-0.03) 0.036 0.01 (0.00-0.02) 0.029 1.39 (0.95-2.02) 0.089 1.41 (0.94-2.10) 0.096 BMI, Body mass index; BMF, Body fat mass, 95% CI, 95% confidence interval; OR, Odds ratio Model 1: Adjusted for gender and age Model 2: Adjusted for gender, age, physical activity and energy intake

183 Chapter 4

Table 4 The dietary intake modifies the GRS association with BFM (%) and the obesity risk (as % of BFM)

p value for p value for β (SE) p value OR (95% CI) p value interaction interaction Energy intake (kcal) ≤7 risk alleles 0.17 (0.05) <0.001 1.00 (1.00-1.00) 0.041 >7 risk alleles 0.39 (0.04) <0.001 0.002* 1.00 (1.00-1.00) <0.001 <0.001* Total protein (g) ≤7 risk alleles 0.11 (0.06) 0.060 1.01 (0.99-1.02) 0.302 >7 risk alleles 0.20 (0.04) <0.001 0.003* 1.02 (1.00-1.03) 0.037 0.001* Animal protein (g) ≤7 risk alleles 0.09 (0.05) 0.053 1.01 (0.99-1.02) 0.261 >7 risk alleles 0.13 (0.04) 0.002 0.032 1.01 (1.00-1.03) 0.080 0.021* Vegetable protein (g) ≤7 risk alleles -0.06 (0.06) 0.340 0.98 (0.94-1.03) 0.484 >7 risk alleles 0.02 (0.06) 0.730 0.003* 1.01 (0.96-1.06) 0.657 0.001* Total fat (g) ≤7 risk alleles 0.05 (0.05) 0.349 1.01 (0.99-1.02) 0.326 >7 risk alleles 0.07 (0.05) 0.179 0.029 1.01 (0.99-1.02) 0.324 0.017* SFA (g) ≤7 risk alleles -0.00 (0.06) 0.938 1.00 (0.93-1.07) 0.945 >7 risk alleles 0.03 (0.05) 0.537 0.015* 0.99 (0.93-1.05) 0.810 0.017* MUFA (g) ≤7 risk alleles 0.07 (0.05) 0.139 1.02 (0.99-1.05) 0.144 >7 risk alleles 0.07 (0.04) 0.123 0.121 1.02 (0.99-1.04) 0.177 0.082 PUFA (g) ≤7 risk alleles -0.05 (0.06) 0.388 0.96 (0.87-1.05) 0.365 >7 risk alleles 0.04 (0.05) 0.396 0.002* 1.06 (0.95-1.18) 0.279 0.004* Diet cholesterol (mg) ≤7 risk alleles 0.09 (0.05) 0.061 1.00 (1.00-1.00() 0.492 >7 risk alleles 0.06 (0.04) 0.163 0.309 1.00 (1.00-1.00) 0.492 0.090 Total carbohydrates (g) ≤7 risk alleles -0.03 (0.05) 0.638 1.00 (0.99-1.00) 0.461 >7 risk alleles 0.05 (0.05) 0.318 0.008* 1.00 (1.00-1.01) 0.402 0.001* Simple carbohydrates (g) ≤7 risk alleles -0.17 (0.05) <0.001 0.99 (0.98-1.00) 0.005 >7 risk alleles -0.16 (0.04) <0.001 0.270 0.97 (0.96-0.98) <0.001 0.997 Complex carbohydrates (g) ≤7 risk alleles 0.05 (0.05) 0.387 1.00 (0.99-1.01) 0.559 >7 risk alleles 0.08 (0.05) 0.153 0.030 1.01 (1.00-1.01) 0.049 0.001* Fiber (g) ≤7 risk alleles -0.09 (0.05) 0.101 0.97 (0.93-1.01) 0.188 >7 risk alleles -0.06 (0.05) 0.245 0.039 0.96 (0.92-1.00) 0.048 0.059 SFA, Saturated fatty acids; MUFA, Monounsaturated fatty acids; PUFA, Polyunsaturated fatty acid; SE, Standard Error; 95% CI, 95% confidence interval Adjusted for gender, age, physical activity and energy intake *p-value <0.05 Benjamini–Hochberg correction for multiple comparisons

184 Results

Supplementary Table 1 Risk alleles, alleles frequency in Hap-Map CEU populations and odds ratio for the association between each SNP and obesity risk

Risk Frequency alleles in Hap-Map Gene SNP Genotype OR (95% CI) p value allele CEU population FTO rs9939609 A TT 1 (ref.) 25.7 TA 1.14 (0.78-1.66) 0.500 56.6 AA 2.08 (1.30-2.30) 0.002 17.7 MC4R rs17782313 C TT 1 (ref.) 50.4 TC 1.04 (0.73-1.47) 0.829 46.0 CC 0.97 (0.43-2.22) 0.951 3.5 MTHFR rs1801133 T CC 1 (ref.) 46.9 CT 1.17 (0.82-1.67) 0.394 44.2 TT 1.06 (0.65-1.73) 0.821 8.8 PPARA rs1800206 G CC 1 (ref.) 91.7 CG 1.01 (0.65-1.56) 0.967 8.2 GG - - 0.0 PPARG rs1801282 G CC 1 (ref.) 82.3 CG 0.91 (0.58-1.42) 0.678 15.9 GG 0.43 (0.07-2.43) 0.339 1.8 APOA5 rs662799 C TT 1 (ref.) 96.7 TC 0.97 (0.59-1.60) 0.918 3.3 CC 2.49 (0.19-32.08) 0.484 0.0 APOE rs429358 C TT 1 (ref.) NA TC 1.49 (0.99-2.26) 0.057 NA CC 0.54 (0.11-2.79) 0.465 NA APOE rs7412 T CC 1 (ref.) NA CT 0.83 (0.50-1.37) 0.460 NA TT 2.77 (0.22-34.04) 0.426 NA LIPC rs1800588 T CC 1 (ref.) 56.7 CT 0.81 (0.57-1.14) 0.222 35.0 TT 1.09 (0.56-2.14) 0.796 8.3 PLIN1 rs894160 A GG 1 (ref.) 46.4 GA 0.87 (0.62-1.23) 0.432 42.9 AA 1.62 (0.86-3.06) 0.132 10.7 NOS3 rs1799983 T GG 1 (ref.) 40.0 GT 1.11 (0.79-1.58) 0.541 51.7 TT 0.73 (0.44-1.24) 0.248 8.3 GCKR rs1260326 T CC 1 (ref.) 33.0 CT 0.89 (0.61-1.29) 0.533 50.0 TT 1.08 (0.68-1.76) 0.755 17.0 LPL rs328 G CC 1 (ref.) 76.7 CG 0.89 (0.62-1.29) 0.553 21.7 GG 0.47 (0.12-1.91) 0.292 1.7 CELSR2 rs12740374 T GG 1 (ref.) 50.0 GT 0.93 (0.65-1.31) 0.661 39.7 TT 1.15 (0.52-2.57) 0.730 10.3 CETP rs1800777 A GG 1 (ref.) 90.0 GA 0.73 (0.31-1.70) 0.465 10.0 AA 1.13 (0.07-18.48) 0.931 0.0 LIPG rs4939883 T CC 1 (ref.) 62.7 CT 1.05 (0.72-1.53) 0.799 35.6 TT 0.88 (0.34-2.25) 0.791 1.7 NA, No available (data) Adjusted for age, sex, physical activity and energy intake

185 Chapter 4

35

p= 0.002 30 25 BFM (%) BFM 20 15 1000 2000 3000 4000

Total energy (kcal) '

Low GRS ( £7 risk alleles) High GRS (>7 risk alleles)

Supplementary Figure 1 Interaction between GRS and total energy intake on percentage of

BFM, after adjusted for age, gender and physical activity. GRS, Genetic risk score; BFM, Body fat mass

186 Results

A 30

p= 0.003 25 BFM (%) BFM 20 15 0 50 100 150 200

Total protein (g) '

Low GRS (£7 risk alleles) High GRS (>7 risk alleles)

B 30

p= 0.032 28 26 24 BFM (%) BFM 22 20

0 50 100 150 200

Animal protein (g) '

Low GRS (£7 risk alleles) High GRS (>7 risk alleles)

C 25 24

p= 0.003 23 22 BFM (%) BFM 21 20

10 20 30 40 50 60

Vegetable protein (g) '

Low GRS (£7 risk alleles) High GRS (>7 risk alleles)

Supplementary Figure 2 Interaction between GRS and A total protein, B vegetable protein and C animal protein intake on percentage of BFM, after adjusted for age, gender, energy intake and physical activity. GRS, Genetic risk score; BFM, Body fat mass

187 Chapter 4

A 26

p= 0.029 24 BFM (%) BFM 22 20

0 50 100 150 200

Total fat (g) '

Low GRS (£7 risk alleles) High GRS (>7 risk alleles)

B 26 25 24 p= 0.015 BFM (%) BFM 23 22

10 20 30 40 50

SFA (g) '

Low GRS (£7 risk alleles) High GRS (>7 risk alleles)

C 26

24 p= 0.002 22 BFM (%) BFM 20 18 0 10 20 30 40

PUFA (g) '

Low GRS (£7 risk alleles) High GRS (>7 risk alleles)

Supplementary Figure 3. Interaction between GRS and A total fat, B SFA and C PUFA intake on percentage of BFM, after adjusted for age, gender, energy intake and physical activity. GRS, Genetic risk score; BFM, Body fat mass; SFA, Saturated fatty acids; PUFA, Polyunsaturated fatty acids

188 Results

A 26 25 24 BFM (%) BFM

23 p= 0.008 22 21 0 100 200 300 400 500

Total carbohydrates (g) '

Low GRS (£7 risk alleles) High GRS (>7 risk alleles)

B 26 25 p= 0.030 24 BFM (%) BFM 23 22 21

0 100 200 300

Complex carbohydrates (g) '

Low GRS (£7 risk alleles) High GRS (>7 risk alleles)

C 26 24 22

BFM (%) BFM p= 0.039 20 18 10 20 30 40 50 60

Fiber (g) '

Low GRS (£7 risk alleles) High GRS (>7 risk alleles)

Supplementary Figure 4 Interaction between GRS and A total carbohydrates, B complex carbohydrates and C fiber intake on percentage of BFM, after adjusted for age, gender, energy intake and physical activity. GRS, Genetic risk score; BFM, Body fat mass

189

CHAPTER 5

Significant phenotype and genotype predictors of BMI in an adult population

Goni L.1,2, García-Granero M.3, Milagro F.I.1,2,4,5, Cuervo M.1,2,4,5, Martínez J.A.1,2,4,5

1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Navarra, Spain 2 Centre for Nutrition Research, University of Navarra, Navarra, Spain 3 Department of Biochemistry and Genetics, University of Navarra, Navarra, Spain 4 Navarra Institute for Health Research (IdiSNA), Navarra, Spain 5 Biomedical Research Centre Network in Physiopathology of Obesity and Nutrition (CIBERobn), Institute of Health Carlos III, Madrid, Spain

Genes Nutr Under review (GNNU-D-17-00050) Impact factor (2016): 2.797 34/81 Nutrition & Dietetics, Q2 75/166 Genetics & Heredity, Q2

Results

ABSTRACT

Obesity is a complex and multifactorial disease resulting from the interactions among genetics, metabolic, behavioral, sociocultural and environmental factors. In this sense, the aim of the present study was to identify phenotype and genotype variables that could be relevant determinants concerning body mass index (BMI) variability. Least angle regression (LARS) analysis was used as regression model selection technique, where the dependent variable was BMI and the independent variables were age, sex, energy intake, physical activity level and 16 polymorphisms previously related to obesity and lipid metabolism. The LARS analysis obtained the following formula for BMI explanation: (64.7 + 0.10 x age [years] + 0.42 x gender [0, men; 1, women] - 40.6 x physical activity [physical activity level] + 0.004 x energy intake [kcal] + 0.74 x rs9939609 [0 or 1-2 risk alleles] + -0.72 x rs1800206 [0 or 1-2 risk alleles] + -0.86 x rs1801282 [0 or 1-2 risk alleles] + 0.87 x rs429358 [0 or 1-2 risk alleles]. The multivariable regression model accounted for 21% of the phenotypic variance in BMI. The regression model was internally validated by the bootstrap method (R2 original dataset=0.208, mean R2 bootstrap datasets=0.210). In conclusion, age, physical activity, energy intake and polymorphisms in FTO, APOE, PPARG and PPARA genes are significant predictors of the BMI trait.

Keywords Genetics, LARS analysis, Multivariable regression model, Obesity

193 Chapter 5

1. INTRODUCTION

In the past 50 years, the prevalence of obesity has steadily raised becoming a global public health problem contributing for a huge increase of health-care costs (Wang et al. 2011). It has been estimated that 2.16 billion adults will be overweight and 1.12 billion adults will be obese by 2030, if the present trends continue (Kelly et al. 2008). An increase in the global burden of overweight and obesity will translate into increase of the risk of several other health conditions including type 2 diabetes, cardiovascular disease or certain types of cancers (Wang et al. 2011). Although obesity it is generally attributed to an imbalance between the energy consumed and the energy expenditure, it is also accepted that it is a complex and multifactorial disease resulting from genetic, physiological, behavioral, sociocultural and environmental factors (Weaver, 2008; Forouzanfar et al. 2015; Kim et al. 2017; Singh et al. 2017; St-Onge 2017;).

Heritability studies indicate that genetic factors could account for 31-90% of the body inter- individual weight variability (Min et al. 2013). However, the large number of single nucleotide polymorphisms (SNPs) identified by genome-wide association studies (GWAS) and candidate gene studies, appeared to explain only 2-4% of the obesity status (El-Sayed Moustafa and Froguel 2013). Even taken together such polymorphisms, they seemed to provide very little risk prediction of the disease (Belsky et al. 2013). In one of the last GWAS related to adiposity, the 97 genome-wide significant loci identified associated with obesity accounted for 2.7% of the body mass index (BMI) variance (Locke et al. 2015).

In addition, a limited predictive value of genetic markers have been described specifically when they are compared to classical non-genetic risk factors (Li et al. 2010; Loos and Janssens 2017). In this context, the design and development of a multivariable regression model based on phenotype and genotype variables could lead us toward the development of more effective precision preventive and treatment dietary interventions (Goni et al. 2016b). Therefore, the aim of the present study was to identify, in an adult population, phenotype and genotype variables, that combined in a multivariable model, could be determinant factors for explaining BMI status.

194 Results

2. MATERIALS AND METHODS

2.1. Study population

The dataset included men and women of Caucasian ancestry, who voluntarily attended community pharmacies in Spain. Genotype information of 1065 individuals was available. Of these, 7 subjects were excluded due to missing values for dietary intake, physical activity and/or anthropometric measurements, and 8 subjects were removed because they were less than 18 y.o. Therefore, a total of 1050 subjects were included in the present study.

Individuals were specifically asked if they would be willing to take part anonymously in the research study. After ensuring that participants had understood the information, only those who provided written informed consent for participation were enrolled. All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation and with the Helsinki Declaration of 1975, as revised in 2000. The Research Ethics Committee of the University of Navarra gave confirmation of fulfillment of the ethical standards and deontological criteria affecting the present survey (Ref. 2410/2014).

2.2. Data collection

Anthropometrics, habitual dietary intake and physical activity measurements were collected by trained nutritionists using a standardized protocol previously described (Goni et al. 2015). Briefly, weight and height were assessed with a digital scale (Tanita BF-522W, Tanita Corporation, Tokyo, Japan) and a portable stadiometer (Leicester Tanita), respectively. BMI was calculated as weight (kg)/height2 (m2).

Habitual dietary intake was determined using a validated food frequency questionnaire, where basic foods were classified into 19 food groups. Subjects were asked to report how often (daily, weekly, monthly or never) they had consumed a choice of each food group. (Goni et al. 2016a). Physical activity was estimated by a short 24 hours physical activity questionnaire in which subjects were asked about the number of hours resting and practicing activities during a weekday and a weekend day (Panel on Macronutrients, Panel on the Definition of Dietary Fiber, Subcommittee on Upper Reference Levels of Nutrients, Subcommittee on Interpretation and Uses of Dietary Reference Intakes 2005).

195 Chapter 5

2.3. DNA isolation and genotyping

Genomic DNA was obtained from oral epithelial cells collected by ORAcollect DNA® (DNAGenotek, Kanata, Ont., Canada). It was isolated by QIAcube using QiAmp DNA Mini QIAcube Kit (Qiagen, Hilden, Germany), following the manufacturer’s procedures. Sixteen polymorphisms previously associated in the scientific database with body weight regulation and lipid metabolism (rs9939609 (FTO), rs17782313 (MC4R), rs1801133 (MTHFR), rs1800206 (PPARA), rs1801282 (PPARG), rs662799 (APOA5), rs429358 (APOE), rs7412 (APOE), rs1800588 (LIPC), rs894160 (PLIN1), rs1799983 (NOS3), rs1260326 (GCKR), rs328 (LPL), rs12740374 (CELSR2), rs1800777 (CETP) and rs4939883 (LIPG)) were genotyped using Luminex® 100/200TM System (Luminex Corporation, Austin, Texas), which is based on the principles of xMAP® Technology (Tai et al. 2002; Chrysohoou et al. 2004; Bennet et al. 2007; Frayling et al. 2007; Loos et al. 2008; Lu et al. 2008; Willer et al. 2008, 2009; Kathiresan et al. 2008; Sandhu et al. 2008; Lewis et al. 2008; Soenen et al. 2009; Ferguson et al. 2010; Kettunen et al. 2012; Galbete et al. 2013). This method uncompressed polystyrene microspheres internally dyed with various ratios of spectrally distinct fluorophores, which are detected by a flow cytometry- based instrument (Dunbar 2006).

2.4. Statistical analyses

Deviation from Hardy-Weinberg equilibrium (HWE) was tested using χ2 test and allele frequencies were estimated. Least angle regression (LARS) analysis was used as regression model selection technique due to its advantages in speed, interpretability and predictive accuracy (Zhang and Zamar 2014). In the current study, the dependent variable was BMI. The independent variables were age, sex, energy intake, physical activity level and the 16 selected polymorphisms. Because LARS algorithm is designed for linear regression with continuous or binary covariates, polymorphisms were recoded in binary variables according to the association between each polymorphism and BMI tested by using dummy linear regression models. In those cases, where there was no significant association and due to the limited frequency of the variant allele, homozygotes of the minor allele (aa) and heterozygotes (Aa) were grouped and compared with major allele homozygotes (AA). Stagewise regression and Lasso were also performed to verify the selection of the independent variables established by LARS (Zhang and Zamar 2014). The independent variables selected by LARS method were combined to generate the regression function. To test potential gene-gene and gene- phenotype interactions among the factors selected by LARS, genotype-by-genotype and

196 Results genotype-by-phenotype product terms were included in the model. Bootstrapping was performed to internally validate the regression model. It was implemented by constructing a number of resamples (K=1000) of the dataset that was obtained by random sampling with replacement from the original dataset. For multiple comparisons Benjamini-Hochberg correction was applied. Statistical analyses were performed using Stata SE, version 12.1 (StataCorp, College Station, TX, USA) and R, version 3.3.2 (R Foundation for Statistical Computing, Vienna, Austria). A p value of p<0.05 was considered as statistically significant.

3. RESULTS

Baseline phenotypic characteristics of the individuals according to gender have been described (Table 1). Although statistically differences were observed for height and weight, no differences were found for BMI. The distribution of the nutritional status differed by gender. Men presented higher percentage of overweight and obese subjects and a less percentage of normal weight individuals than women. As expected, there were statistically significant differences for energy intake and physical activity level, being the consumption of energy and the physical activity practice higher among males than among females.

The genotypes distribution, minor allele frequencies (MAF) and HWE for each polymorphism are listed (Table 2). MAF ranged from 0.02-0.45. The distributions of all the polymorphisms alleles were in HWE except the rs1800588 polymorphism located in LIPC gene even after Benjamini-Hochberg correction for multiple comparisons.

According to the LARS analysis age, physical activity, energy intake and 4 polymorphisms were involved in BMI explanation (Table 3 and Figure 1). Although gender was not selected by LARS it was included in the model as a common cofounding factor. The LARS analysis obtained the following formula for explain BMI: (64.7 + 0.10 x age [years] + 0.42 x gender [0, men; 1, women] - 40.6 x physical activity [physical activity level] + 0.004 x energy intake [kcal] + 0.74 x rs9939609 [0 or 1-2 risk alleles] + -0.72 x rs1800206 [0 or 1-2 risk alleles] + -0.86 x rs1801282 [0 or 1-2 risk alleles] + 0.87 x rs429358 [0 or 1-2 risk alleles]. The multivariable regression model accounted for 21% of the phenotypic variance in BMI. The selection of the independent variables established by LARS was confirmed by stagewise regression and Lasso (data not shown).

Additionally, gene-phenotypic factors and gene-gene interactions were tested. Trend toward significance interactions were found for FTO polymorphism and energy intake and for PPARA genetic variant and energy intake. When both product terms of the interactions were included

197 Chapter 5 in the regression model the adjusted r squared did not improve significantly (adjusted R2 for regression model 0.208; adjusted R2 for the regression model including interactions 0.212).

In order to evaluate the accuracy of the model, the relationship between the observed and the predicted BMI was plotted (Figure 2). The predicted BMI agrees with the observed or “real” BMI by checking the parameters of the linear regression. The intercept of the model is very close to zero and the slope is almost 1, meaning that the change in both variables can be considered proportional.

The internal validation was performed by the bootstrap method, whose estimates agreed closely with the parameters obtained by LARS (R2 original dataset=0.208, mean R2 bootstrap datasets=0.210).

4. DISCUSSION

Because common obesity is a multifactorial disease, where genetic, metabolic, physiological, behavioral, sociocultural and environmental factors are involved, in the current study a regression model based on phenotype and genotype determinants of BMI has been defined. The regression model includes a total of 4 phenotypic characteristics (age, gender, energy intake and physical activity) and 4 polymorphisms located next to or in FTO, APOE, PPARG, and PPARA genes.

The LARS analysis reported 4 polymorphisms significantly or marginally associated with BMI located in FTO, APOE, PPARG and PPARA genes. The rs9939609 FTO variant is the most well- known polymorphism associated with obesity. FTO is a nuclear protein, which is a member of the AlkB related non-haem iron and 2-oxoglutarate-dependent oxygenase superfamily (Gerken et al. 2007). Although the relationship between FTO genetic variant and obesity related traits (BMI, obesity risk, waist circumference, body fat mass) has been confirmed in several populations, the physiological function of this gene in body weight regulation seems unclear (Zhao et al. 2014). Results from animal and human studies suggest that FTO gene acts on food intake with no impact on resting energy expenditure, although in rodents results are controversial (Speakman 2015). Since Fto in rodents is widely expressed in the brain, including the hypothalamic nucleic, it has been proposed that FTO could be linked to BDNF-NTRK2 signaling pathway influencing food intake regulation (Rask-Andersen et al. 2011). Moreover, Fto has been implicated in the regulation of leptin levels and leptin sensitivity (Wang and Yang 2011; Kilpeläinen et al. 2016). Additionally, SNPs in FTO could exerts their effects in other genes including RLGRIPIL and IRX3 (Stratigopoulos et al. 2014; Smemo et al. 2014).

198 Results

As far as we know our group reported for the first time a significant association between rs429358 APOE genetic variant and BMI (Goni et al. 2015). In the present study, such relationship has been verified in a large sample. The APOE gene encodes for a structural apolipoprotein that plays a major role in maintaining plasma lipids homeostasis (Hatters et al. 2006). APOE genetic variants have been associated with several metabolic disorders including high obesity risk (Volcik et al. 2006; Zeljko et al. 2011; Petkeviciene et al. 2012; Mazzotti et al. 2014). The contribution of APOE to obesity could be explained because of its effect on adipogenesis as has been reported in vivo and in vitro studies (Kypreos et al. 2009; Li and Liu 2014).

PPARG and PPARA are members of the nuclear hormone receptor subfamily of ligand- dependent transcription factors. PPARG is primarily expressed in adipocytes where it modulates the expression of target genes involved in adipocyte differentiation, insulin sensitivity and inflammatory processes (Fajas et al. 1997; Robinson and Grieve 2009). In contrast, PPARA is expressed in liver, heart, kidney and skeletal muscle to regulate fatty acid oxidation systems (Robinson and Grieve 2009). The Pro12Ala (rs1801282), which is the PPARG genetic variant most studied in relation to adiposity and insulin resistance, has been associated with reduced ability to transactivate responsive promoters and with lower transcriptional activity (Deeb et al. 1998; Gouda et al. 2010; Galbete et al. 2013). Although in most of candidate gene studies Pro12Ala has been associated with higher BMI, other authors reported the opposite association or have not found any association at all (Deeb et al. 1998; Evans et al. 2000; Galbete et al. 2013; Baldani et al. 2014; Hsiao and Lin 2015; Mansoori et al. 2015). These controversial results suggest that, if this variant does influence obesity predisposition, it may do so through environment-dependent mechanisms. In fact, several studies have reported interactions between PPARG and environmental factors such as gender, dietary fat intake or breast feeding on obesity traits (Verier et al. 2010; Dedoussis et al. 2011; Lamri et al. 2012; Randall et al. 2013). Although the association between genetic variants of the PPARG gene and obesity traits has been widely studied, as far as we know there is limited evidence regarding the relationship between PPARA variants and obesity phenotype. Meanwhile Costa-Urrutia et al. (2017) reported a positive association between the rs1800206 PPARA polymorphism and obesity risk; Sirbelnagel et al. (2009) did not find a relationship between such genetic variant and BMI or body fat composition (Silbernagel et al. 2009; Costa-Urrutia et al. 2017). We hypothesized that our opposite results regarding PPARG and PPARA could be due partly to the fact that we have carried out the analysis in the presence of other genetic variants.

199 Chapter 5

Interestingly, 21% of the phenotypic variance in BMI was accounted using the regression model obtained by LARS including gender, age, energy intake, physical activity and four genetic variants located near or in FTO, APOE, PPARG and PPARA genes. The four polymorphisms accounted for 0.5% of the BMI variability. This finding is in accordance with the studies by Martínez-García et al. (2013), Belsky et al. (2013) and Li et al. (2010), in which a small number of SNPs explained less than 1% of the BMI heritability (Martínez-García et al. 2013; Belsky et al. 2013; Li et al. 2010). In this sense, it should be highlighted that when Locke et al. (2015) included a total of 97 SNPs in a prediction model of BMI the authors found a BMI explanation of 2.7% (Locke et al. 2015). As far as we know, prediction models that added energy intake and physical activity have not been reported up to date, so we cannot be able to compare our results. However, some authors have observed that, when phenotypical factors are included in the genetic model (such as socioeconomic or depression status), the percentage of the explanation of the BMI significantly increases (Belsky et al. 2013; Hung et al. 2015).

Several potential explanations can be offered for the low predictive value of the regression model, but are mainly related to the fact that obesity is characterized for being a multifactorial disease. Although we have included in the model the two main factors that characterized obesity, energy intake and physical activity, there are other features that have not been taken into account such as social determinants (education level, economic status), endocrine disorders (hypothyroidism) or use of certain medications (Weaver, 2008; El-Sayed et al. 2012; Martínez de Morentin-Aldabe et al. 2013; Kim et al. 2017). Another explanation for the low predictive value of the regression model could be related with the marginal effect sizes of the tested variants and the skewed distribution of the effect sizes. In addition, predictive models could include other sources of variation known or hypothesized to influence BMI, such as rare variants, gene-gene and gene-environment interactions, copy number variation and epigenetic and metagenomic effects (Goni et al. 2016b). Finally, it should be mentioned that in the present study BMI instead of body fat mass was selected as dependent variable. Although BMI is the adiposity measurement most widely used in epidemiological studies, its interpretation does not differ between gender and race, and neither distinguishes between degree of fatness, muscle mass and skeletal mass (Gómez-Ambrosi et al. 2012). Therefore, it can lead to errors in the estimation of adiposity, over or underestimating adiposity depending on subject complexion; such as athletes or metabolic obese normal weight individuals.

To the best of our knowledge, this is the first study that applies LARS analysis to select phenotype and genotype variables for explain BMI status. However, the study bears some limitations that need to be mentioned. First, the present study included only subjects of

200 Results

Caucasian ancestry, so the findings may not be generalizable to other ethnic groups. Second, the model developed in this study used BMI as the response variable instead of body fat mass. Although BMI has some limitations in its interpretability, it is the adiposity measurement more used in epidemiological studies.

In conclusion, significant predictors of BMI included age, energy intake, physical activity and polymorphisms located near or in FTO (rs9939609), APOE (rs429358), PPARG (rs1801282) and PPARA (rs1800206). Although 4 polymorphisms were selected by LARS, it should be mentioned that they explain a small percentage of BMI variation as have found other authors. Moreover, the proposed statistical method, LARS analysis, could help to implement new criteria for the identification of BMI predictors since obesity is a multifactorial disease in which a large number of phenotypic and genotypic features are involved.

DECLARATION Ethics approval and consent to participate All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The Research Ethics Committee of the University of Navarra gave confirmation of fulfillment of the ethical standards and deontological criteria affecting the present survey (Ref. 2410/2014). Informed consent was obtained from all individual participants included in the study.

Consent for publication Not applicable

Availability of data and material No additional data are available.

Competing interests Leticia Goni, Marta García-Granero, Fermín I. Milagro Marta Cuervo, and J. Alfredo Martínez declare that they have no conflict of interest.

201 Chapter 5

Funding Linea Especial (University of Navarra; LE/97). Spanish Ministry of Economy and Competitiveness (AGL2013-45554-R project). CIBERobn/RETICS schedules (Institute of Health Carlos III).

Author’s contributions LG, MG-G and MC contributed to conception and designed the study. LG and MG-G contributed to analysis and interpretation of data. LG and JAM contributed to drafted the manuscript. MG-G, FIM, MC and JAM contributed to critical revision of the manuscript.

Acknowledgments The authors thank all the users of the Nutrigenetic Service who voluntarily offer her data to the study. The authors are grateful to the nutritionists for data collection and to Amaia Ibañez for excellent technical assistance. Moreover, we wish to thank Enrique Goñi for excellent statistical assistance in the University of Navarra. The predoctoral research grant to Leticia Goni from the Spanish Ministry of Education, Culture and Sport is gratefully acknowledged. The authors also wish to thank the support from CINFA concerning the genetic tools and general logistic.

202 Results

REFERENCES

Baldani DP, Skrgatic L, Cerne JZ, et al (2014) Association of PPARG Pro12Ala polymorphism with insulin sensitivity and body mass index in patients with polycystic ovary syndrome. Biomed Reports 2:199–206. doi: 10.3892/br.2013.215 Belsky DW, Moffitt TE, Sugden K, et al (2013) Development and evaluation of a genetic risk score for obesity. Biodemography Soc Biol 59:85–100. doi: 10.1080/19485565.2013.774628; 10.1080/19485565.2013.774628 Bennet AM, Di Angelantonio E, Ye Z, et al (2007) Association of apolipoprotein E genotypes with lipid levels and coronary risk. JAMA 298:1300–1311. doi: 10.1001/jama.298.11.1300 Chrysohoou C, Panagiotakos DB, Pitsavos C, et al (2004) Evidence for association between endothelial nitric oxide synthase gene polymorphism (G894T) and inflammatory markers: the ATTICA study. Am Heart J 148:733–738. doi: 10.1016/j.ahj.2004.04.022 Costa-Urrutia P, Abud C, Franco-Trecu V, et al (2017) Genetic obesity risk and attenuation effect of physical fitness in Mexican-Mestizo population: a case-control study. Ann Hum Genet 81:106–116. doi: 10.1111/ahg.12190 Dedoussis G V, Manios Y, Kourlaba G, et al (2011) An age-dependent diet-modified effect of the PPARgamma Pro12Ala polymorphism in children. Metabolism 60:467–473. doi: 10.1016/j.metabol.2010.04.007; 10.1016/j.metabol.2010.04.007 Deeb SS, Fajas L, Nemoto M, et al (1998) A Pro12Ala substitution in PPARgamma2 associated with decreased receptor activity, lower body mass index and improved insulin sensitivity. Nat Genet 20:284–7. doi: 10.1038/3099 Dunbar SA (2006) Applications of Luminex xMAP technology for rapid, high-throughput multiplexed nucleic acid detection. Clin Chim Acta 363:71–82. doi: 10.1016/j.cccn.2005.06.023 El-Sayed AM, Scarborough P, Galea S (2012) Unevenly distributed: a systematic review of the health literature about socioeconomic inequalities in adult obesity in the United Kingdom. BMC Public Health 12:18–30. doi: 10.1186/1471- 2458-12-18 El-Sayed Moustafa JS, Froguel P (2013) From obesity genetics to the future of personalized obesity therapy. Nat Rev 9:402–413. doi: 10.1038/nrendo.2013.57 [doi] Evans D, Mann WA, de Heer J, et al (2000) Variation in the gene for human peroxisome proliferator activated receptor gamma (PPARgamma) does not play a major role in the development of morbid obesity 2. IntJ Obes Relat Metab Disord 24:647–651. doi: 10.1038/sj.ijo.0801214 Fajas L, Auboeuf D, Raspé E, et al (1997) The organization, promoter analysis, and expression of the human PPARgamma gene. J Biol Chem 272:18779–18789. doi: 10.1074/jbc.272.30.18779 Ferguson JF, Phillips CM, McMonagle J, et al (2010) NOS3 gene polymorphisms are associated with risk markers of cardiovascular disease, and interact with omega-3 polyunsaturated fatty acids. Atherosclerosis 211:539–544. doi: 10.1016/j.atherosclerosis.2010.03.027; 10.1016/j.atherosclerosis.2010.03.027 Forouzanfar MH, Alexander L, Anderson HR, et al (2015) Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks in 188 countries, 1990– 2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 386:2287–2323. doi: 10.1016/S0140-6736(15)00128-2 Frayling TM, Timpson NJ, Weedon MN, et al (2007) A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science 316:889–94. doi: 10.1126/science.1141634 Galbete C, Toledo E, Martínez-González MA, et al (2013) Pro12Ala variant of the PPARG2 gene increases body mass index: an updated meta-analysis encompassing 49,092 subjects. Obesity 21:1486–1495. doi: 10.1002/oby.20150

203 Chapter 5

Gerken T, Girard CA, Tung Y-CL, et al (2007) The obesity-associated FTO gene encodes a 2-oxoglutarate-dependent nucleic acid demethylase. Science (80- ) 318:1469–1472. doi: 10.1126/science.1151710 Gómez-Ambrosi J, Silva C, Galofré JC, et al (2012) Body mass index classification misses subjects with increased cardiometabolic risk factors related to elevated adiposity. Int J Obes 36:286–294. doi: 10.1038/ijo.2011.100 Goni L, Aray-Miranda M, Martínez JA, Cuervo M (2016a) Validation of a food groups frequency questionnaire based in an exchange system. Nutr Hosp 33:1391–1399. doi: http://dx.doi.org/10.20960/nh.574 Goni L, Cuervo M, Milagro FI, Mart JA (2016b) Future Perspectives of Personalized Weight Loss Interventions Based on Nutrigenetic, Epigenetic, and Metagenomic Data. J Nutr 146:905–912. doi: 10.3945/jn.115.218354.The Goni L, Cuervo M, Milagro FI, Martinez JA (2015) A genetic risk tool for obesity predisposition assessment and personalized nutrition implementation based on macronutrient intake. Genes Nutr 10:445–014–0445– Epub 2014 Nov 28. doi: 10.1007/s12263-014-0445-z [doi] Gouda HN, Sagoo GS, Harding AH, et al (2010) The association between the peroxisome proliferator-activated receptor-γ2 (PPARG2) Pro12Ala gene variant and type 2 diabetes mellitus: A HuGE review and meta-analysis. Am J Epidemiol 171:645–655. doi: 10.1093/aje/kwp450 Hatters DM, Peters-Libeu CA, Weisgraber KH (2006) Apolipoprotein E structure: insights into function. Trends Biochem Sci 31:445–454. doi: 10.1016/j.tibs.2006.06.008 Hsiao TJ, Lin E (2015) The Pro12Ala polymorphism in the peroxisome proliferator-activated receptor gamma (PPARG) gene in relation to obesity and metabolic phenotypes in a Taiwanese population. Endocrine 48:786–793. doi: 10.1007/s12020-014-0407-7 Hung CF, Breen G, Czamara D, et al (2015) A genetic risk score combining 32 SNPs is associated with body mass index and improves obesity prediction in people with major depressive disorder. BMC Med 13:86. doi: 10.1186/s12916-015-0334-3 Kathiresan S, Melander O, Guiducci C, et al (2008) Six new loci associated with blood low-density lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in humans. Nat Genet 40:189–197. doi: 10.1038/ng.75; 10.1038/ng.75 Kelly T, Yang W, Chen C-S, et al (2008) Global burden of obesity in 2005 and projections to 2030. Int J Obes (Lond) 32:1431–7. doi: 10.1038/ijo.2008.102 Kettunen J, Tukiainen T, Sarin A-P, et al (2012) Genome-wide association study identifies multiple loci influencing human serum metabolite levels. Nat Genet 44:269–76. doi: 10.1038/ng.1073 Kilpeläinen TO, Carli JFM, Skowronski AA, et al (2016) Genome-wide meta-analysis uncovers novel loci influencing circulating leptin levels. Nat Commun 7:10494. doi: 10.1038/ncomms10494 Kim TJ, Roesler NM, von dem Knesebeck O (2017) Causation or selection - examining the relation between education and overweight/obesity in prospective observational studies: a meta-analysis. Obes Rev. doi: 10.1111/obr.12537 Kypreos KE, Karagiannides I, Fotiadou EH, et al (2009) Mechanisms of obesity and related pathologies: role of apolipoprotein E in the development of obesity. FEBS J 276:5720–5728. doi: 10.1111/j.1742-4658.2009.07301.x; 10.1111/j.1742-4658.2009.07301.x Lamri A, Abi Khalil C, Jaziri R, et al (2012) Dietary fat intake and polymorphisms at the PPARG locus modulate BMI and type 2 diabetes risk in the D.E.S.I.R. prospective study. Int J Obes 36:218–224. doi: 10.1038/ijo.2011.91 Lewis SJ, Lawlor DA, Nordestgaard BG, et al (2008) The methylenetetrahydrofolate reductase C677T genotype and the risk of obesity in three large population-based cohorts. Eur J Endocrinol 159:35–40. doi: 10.1530/EJE-08-0056; 10.1530/EJE-08-0056

204 Results

Li S, Zhao JH, Luan J, et al (2010) Cumulative effects and predictive value of common obesity-susceptibility variants identified by genome-wide association studies. Am J Clin Nutr 91:184–190. doi: 10.3945/ajcn.2009.28403; 10.3945/ajcn.2009.28403 Li Y, Liu L (2014) Apolipoprotein E synthesized by adipocyte and apolipoprotein E carried on lipoproteins modulate adipocyte triglyceride content. Lipids Health Dis 13:136. doi: 10.1186/1476-511X-13-136 Locke AE, Kahali B, Berndt SI, et al (2015) Genetic studies of body mass index yield new insights for obesity biology. Nature 518:197–206. doi: 10.1038/nature14177 Loos RJF, Janssens ACJW (2017) Predicting polygenic obesity using genetic information. Cell Metab 25:535–543. doi: 10.1016/j.cmet.2017.02.013 Loos RJF, Lindgren CM, Li S, et al (2008) Common variants near MC4R are associated with fat mass, weight and risk of obesity. Nat Genet 40:768–775. doi: 10.1038/ng.140 Lu Y, Dolle ME, Imholz S, et al (2008) Multiple genetic variants along candidate pathways influence plasma high- density lipoprotein cholesterol concentrations. J Lipid Res 49:2582–2589. doi: 10.1194/jlr.M800232-JLR200; 10.1194/jlr.M800232-JLR200 Mansoori A, Amini M, Kolahdooz F, Seyedrezazadeh E (2015) Obesity and Pro12Ala polymorphism of peroxisome proliferator-activated receptor-gamma gene in healthy adults: a systematic review and meta-analysis. Ann Nutr Metab 67:104–118. doi: 10.1159/000439285 Martínez-García F, Mansego ML, Rojo-Martínez G, et al (2013) Impact of obesity-related genes in Spanish population. BMC Genet 14:111. doi: 10.1186/1471-2156-14-111 Martínez de Morentin-Aldabe BE, Hernández-Ruiz de Eguilaz M, Pérez-Díez S, Martínez JA (2013) Influencia de la administración de psicofármacos en el aumento del peso corporal. Rev Española Nutr Humana y Dietética 17:17–26. Mazzotti DR, Singulane CC, Ota VK, et al (2014) Association of APOE, GCPII and MMP9 polymorphisms with common diseases and lipid levels in an older adult/elderly cohort. Gene 535:370–375. doi: 10.1016/j.gene.2013.11.040 Min J, Chiu DT, Wang Y (2013) Variation in the heritability of body mass index based on diverse twin studies: a systematic review. Obes Rev 14:871–882. doi: 10.1111/obr.12065 Panel on Macronutrients, Panel on the Definition of Dietary Fiber, Subcommittee on Upper Reference Levels of Nutrients, Sub- committee on Interpretation and Uses of Dietary Reference Intakes and the SC on the SE of DRI (2005) Dietary reference intakes for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein, and amino acids (Macronutrients). Washington Petkeviciene J, Smalinskiene A, Luksiene DI, et al (2012) Associations between apolipoprotein E genotype, diet, body mass index, and serum lipids in Lithuanian adult population. PLoS One 7:2–7. doi: 10.1371/journal.pone.0041525 Randall JC, Winkler TW, Kutalik Z, et al (2013) Sex-stratified genome-wide association studies including 270,000 individuals show sexual dimorphism in genetic loci for anthropometric traits. PLoS Genet 9:e1003500. doi: 10.1371/journal.pgen.1003500; 10.1371/journal.pgen.1003500 Rask-Andersen M, Almén M, Olausen HR, et al (2011) Functional coupling analysis suggests link between the obesity gene FTO and the BDNF-NTRK2 signaling pathway. BMC Neurosci 12:117. doi: 10.1186/1471-2202-12-117 Robinson E, Grieve DJ (2009) Significance of peroxisome proliferator-activated receptors in the cardiovascular system in health and disease. Pharmacol Ther 122:246–63. doi: 10.1016/j.pharmthera.2009.03.003 Sandhu MS, Waterworth DM, Debenham SL, et al (2008) LDL-cholesterol concentrations: a genome-wide association study. Lancet 371:483–491. doi: 10.1016/S0140-6736(08)60208-1; 10.1016/S0140-6736(08)60208-1

205 Chapter 5

Silbernagel G, Stefan N, Hoffmann MM, et al (2009) The L162V polymorphism of the peroxisome proliferator activated receptor alpha gene (PPARA) is not associated with type 2 diabetes, BMI or body fat composition. Exp Clin Endocrinol Diabetes 117:113–118. doi: 10.1055/s-0028-1082069 Singh RK, Kumar P, Mahalingam K (2017) Molecular genetics of human obesity: a comprehensive review. C R Biol 340:87–108. doi: 10.1016/j.crvi.2016.11.007 Smemo S, Tena JJ, Kim KH, et al (2014) Obesity-associated variants within FTO form long-range functional connections with IRX3. Nature 507:371–375. doi: 10.1038/nature13138 [doi] Soenen S, Mariman EC, Vogels N, et al (2009) Relationship between perilipin gene polymorphisms and body weight and body composition during weight loss and weight maintenance. Physiol Behav 96:723–728. Speakman JR (2015) The “Fat Mass and Obesity Related” (FTO) gene: Mechanisms of impact on obesity and energy balance. Curr Obes Rep 4:73–91. doi: 10.1007/s13679-015-0143-1 St-Onge M-P (2017) Sleep-obesity relation: underlying mechanisms and consequences for treatment. Obes Rev 18:34–39. doi: 10.1111/obr.12499 Stratigopoulos G, Martin Carli JF, O’Day DR, et al (2014) Hypomorphism for RPGRIP1L, a ciliary gene vicinal to the fto locus, causes increased adiposity in mice. Cell Metab 19:767–779. doi: 10.1016/j.cmet.2014.04.009 Tai ES, Demissie S, Cupples LA, et al (2002) Association between the PPARA L162V polymorphism and plasma lipid levels: the Framingham Offspring Study. Arterioscler Thromb Vasc Biol 22:805–810. Verier MD C, Meirhaeghe PHD A, Bokor MD, PHD S, et al (2010) Breast-feeding modulates the influence of the Peroxisome Proliferator-Activated Receptor-[gamma] (PPARG) Pro12Ala polymorphism on adiposity in adolescents. Diabetes Care 33:190–196. doi: 10.2337/dc09-1459. Volcik KA, Barkley RA, Hutchinson RG, et al (2006) Apolipoprotein E polymorphisms predict low density lipoprotein cholesterol levels and carotid artery wall thickness but not incident coronary heart disease in 12,491 ARIC study participants. Am J Epidemiol 164:342–348. doi: 10.1093/aje/kwj202 Wang P, Yang F-J (2011) Involvement of leptin receptor long isoform (LepRb)-STAT3 signaling pathway in brain Fat Mass– and Obesity-Associated (FTO) downregulation during energy restriction. Mol Med 17:1. doi: 10.2119/molmed.2010.000134 Wang YC, McPherson K, Marsh T, et al (2011) Health and economic burden of the projected obesity trends in the USA and the UK. Lancet 378:815–825. doi: 10.1016/S0140-6736(11)60814-3; 10.1016/S0140-6736(11)60814-3 Weaver JU (2008) Classical endocrine diseases causing obesity. Front Horm Res 36:212–28. doi: 10.1159/0000115367 Willer CJ, Sanna S, Jackson AU, et al (2008) Newly identified loci that influence lipid concentrations and risk of coronary artery disease. Nat Genet 40:161–169. doi: 10.1038/ng.76; 10.1038/ng.76 Willer CJ, Speliotes EK, Loos RJF, et al (2009) Six new loci associated with body mass index highlight a neuronal influence on body weight regulation. Nat Genet 41:25–34. doi: Doi 10.1038/Ng.287 Zeljko HM, Škarić-Jurić T, Narančić NS, et al (2011) E2 allele of the apolipoprotein E gene polymorphism is predictive for obesity status in Roma minority population of Croatia. Lipids Health Dis 10:9–14. doi: 10.1186/1476-511X-10-9 Zhang H, Zamar RH (2014) Least angle regression for model selection. WIREs Comput Stat 6:116–123. Zhao X, Yang Y, Sun BF, et al (2014) FTO and obesity: Mechanisms of association. Curr Diab Rep 14:486–495. doi: 10.1007/s11892-014-0486-0

206 Results

Table 1. Anthropometrical and nutritional characteristics

Total population Males Females p (n=1050) (n=252) (n=798) Age (years) 50.0 (12.8) 47.4 (12.9) 50.8 (12.7) <0.001 Height (cm) 164 (9.0) 174 (7.2) 160 (6.6) <0.001 Weight (kg) 79.5 (17.1) 91.4 (16.4) 75.7 (15.5) <0.001 BMI (kg/m2) 29.6 (5.9) 29.9 (5.0) 29.4 (6.1) 0.23 Nutritional status 0.001 Normal weight (BMI 18.5-24.9 kg/m2) 232 (22.1) 32 (12.7) 200 (25.1) Overweight (BMI 25-29.9 kg/m2) 377 (35.9) 106 (42.1) 271 (34.0) Obesity (BMI ≥30kg/m2) (%) 441 (42.0) 114 (45.2) 327 (41.0) Energy intake (kcal) 2138 (453) 2438 (492) 2043 (395) <0.001 Physical activity 1.23 (0.03) 1.24 (0.03) 1.23 (0.03) <0.001 level BMI, Body mass index T-test for continuous variables and expressed as medium (standard deviation); and Chi-squared test for qualitative variables and expressed as n (%)

207 Chapter 5

Table 2. Genotype, minor allele frequency and Hardy–Weinberg equilibrium calculations of the 16 SNPs included in LARS analysis

Major/minor Major allele Heterozygote Minor allele Gene SNP MAF HWE p-value allele homozygote (%) (%) homozygote (%) FTO rs9939609 T/A 329 (31.3) 519 (49.4) 202 (19.2) 0.44 0.92 MC4R rs17782313 T/C 646 (61.5) 355 (33.8) 49 (4.7) 0.22 0.98

MTHFR rs1801133 C/T 380 (36.2) 511 (48.7) 159 (15.1) 0.39 0.55 PPARA rs1800206 C/G 881 (83.9) 162 (15.4) 7 (0.7) 0.08 0.88

PPARG rs1801282 C/G 889 (84.7) 152 (14.5) 9 (0.9) 0.08 0.38 APOA5 rs662799 T/C 919 (87.5) 126 (12.0) 5 (0.5) 0.06 0.46 APOE rs429358 T/C 842 (80.2) 199 (18.9) 9 (0.9) 0.10 0.76 APOE rs7412 C/T 920 (87.6) 127 (12.1) 3 (0.3) 0.06 0.53 LIPC rs1800588 C/T 617 (58.8) 360 (34.3) 73 (6.9) 0.24 0.04* PLIN1 rs894160 G/A 560 (53.3) 404 (38.5) 86 (8.2) 0.27 0.28 NOS3 rs1799983 G/T 418 (39.8) 488 (46.5) 144 (13.7) 0.37 0.93 GCKR rs1260326 C/T 315 (30.0) 527 (50.2) 208 (19.8) 0.45 0.64 LPL rs328 C/G 765 (72.9) 264 (25.1) 21 (2.0) 0.15 0.75

CELSR2 rs12740374 G/T 658 (62.3) 250 (33.3) 42 (4.0) 0.21 0.59 CETP rs1800777 G/A 1002 (95.4) 46 (4.4) 2 (0.2) 0.02 0.06 LIPG rs4939883 C/T 756 (72.0) 260 (24.8) 34 (3.2) 0.16 0.05

MAF, Minor Allele Frequency; HWE, Hardy Weinberg Equilibrium * p value <0.05 after Benjamini–Hochberg correction for multiple comparisons

208 Results

Table 3. Regression coefficients of the variables selected by LARS

Variables selected b 95% CI p value Energy intake (kcal) 0.004 0.003;0.005 <0.001 Physical activity level -40.6 -49.9;-31.3 <0.001 Age (years) 0.10 0.08;0.13 <0.001 rs429358 (APOE) a 0.86 0.06;1.66 0.035 rs9939609 (FTO) a 0.74 0.05;1.42 0.035 rs1800206 (PPARA) a -0.72 -1.58;0.15 0.104 rs1801282 (PPARG) a -0.86 -1.74;0.043 0.058 Gender 0.42 -0.39;1.24 0.304 Constant 59.1 43.3;74.9 <0.001 a Dominant model

209 Chapter 5

Figure 1. LARS analysis

210 Results

Figure 2. Correlation coefficient between observed BMI and predicted BMI based on the multivariable regression model obtained by LARS

211

CHAPTER 6

Interaction between ADCY3 genetic variant and two weight loss diets: changes in body fatness and body composition

Goni L.1,2, Riezu-Boj J.I.1,2,3, Milagro F.I.1,2,3,4, Corrales F.J.5,6, Ortiz L., Cuervo M.1,2,3,4, Martínez J.A. 1,2,3,4

1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Navarra, Spain 2 Centre for Nutrition Research, University of Navarra, Navarra, Spain 3 Navarra Institute for Health Research (IdiSNA), Navarra, Spain 4 Biomedical Research Centre Network in Physiopathology of Obesity and Nutrition (CIBERobn), Institute of Health Carlos III, Madrid, Spain 5 Centro Nacional de Biotecnología (CNB-CSIC), Madrid, Spain 6 Biomedical Research Centre Network in Hepatic and Gastrointestinal diseases (CIBERehd), Institute of Health Carlos III, Madrid, Spain

PLOS ONE Under review (PONE-D-17-24770) Impact factor (2016): 2.806 15/64 Multidisciplinary Sciences, Q1

Results

ABSTRACT

Adenylate cyclase 3 (ADCY3) gene is involved in a number of physiological processes and metabolic pathways including the development and function of the adipose tissue, since it catalyzes the formation of cyclic adenosine monophosphate (cAMP). Different polymorphisms located near or in the ADCY3 gene have been related to obesity phenotype. The effects of the ADCY3 rs10182181 genetic variant on changes in body fatness and composition depending on the macronutrient distribution intake after 16 weeks of the dietary intervention were tested. The ADCY3 rs10182181 genetic variant was genotyped in 147 overweight or obese subjects, who were randomly assigned to one of the two diets varying in macronutrient content: a moderately-high-protein diet and a low-fat diet. Anthropometric (weight, waist circumference) and body composition (total fat, trunk fat, android fat, gynoid fat and visceral fat) measurements, using a DEXA scan, were recoded. Significant interactions between the ADCY3 genotype and dietary intervention on changes in weight, waist circumference (WC) and body composition after adjusted for covariates (all p for interaction <0.05 in additive pattern) were found. Thus, in the moderately-high-protein diet group, the G allele was associated to a lower decrease of fat mas, trunk fat and android fat, and a greater decrease of lean mass (p=0.03, p=0.03, p=0.01 and p=0.04, respectively). Conversely, in the low-fat diet group carrying the G allele was associated with a greater decrease in trunk fat, android fat, gynoid fat and visceral fat (p=0.04, p=0.02, p=0.03 and p=0.02, respectively). In conclusion, subjects carrying the G allele of the ADCY3 genetic variant may benefit more in weight loss and improvement of body composition measurements when undertaking a hypocaloric low-fat diet as compared to a moderately-high-protein diet.

215 Chapter 6

1. INTRODUCTION

Obesity is a multifactorial disease where various pathophysiological processes are involved including the hormonal regulation of hunger and satiety, the activity of the central reward system, whole-body energy expenditure and the storage capacity for fat in the adipose tissue and the interactions with environmental factors [1]. In each of these metabolic functions a set of genes is implicated with some of them being involved in a broad spectrum of processes [2]. The pleiotropic effect of such genes could be of special interest because they might explain a part of the obesity causes.

Adenylate cyclase 3 (ADCY3), which encodes for a membrane-associated enzyme that catalyzes the formation of the secondary messenger cyclic adenosine monophosphate (cAMP) from adenosine triphosphate (ATP), could be a good candidate gene since it is widely expressed in most cell types and may be involved in a number of physiological and pathophysiological processes [3]. In animal studies, ADCY3 expression has been found in certain regions of the brain, including striatum and hypothalamus, and in the adipose tissue [4,5]. In humans, different genetic variants located near or in ADCY3 gene have been associated with obesity trough candidate gene studies and genome-wide association studies (GWAS) [6–12]. Among the ADCY3 genetic variants identified to be related to obesity traits, the polymorphism rs10182181 has been replicated in the largest meta-analysis of GWAS on body mass index (BMI) carried out up to date [12].

In this context, the aim of the present study was to analyze the effect of two different hypo- energetic diets with varying macronutrient distribution (moderately-high-protein diet vs low- fat diet) on changes in anthropometric and body composition measurements after a 16-week dietary intervention according to the ADCY3 rs10182181 genetic variant.

The present study provides the first gene-diet interaction effect between ADCY3 polymorphisms and moderately-high-protein diet / low-fat diet on changes in body fatness and body composition traits. Our study evidence that gene-diet interaction analyses may give new insights to the development of more precise dietary strategies based on the genotype.

2. MATERIALS AND METHODS

2.1. Study population

The study encompassed a total of 147 overweight or obese subjects (BMI: 25-40 kg/m2) enrolled in a 16-weeks randomized clinical trial (clinical trial reg. no. NCT02737267,

216 Results clinicaltrials.gov) of two hypo-energetic diets with different macronutrient composition, a low- fat diet and a moderately-high-protein diet (Figure 1). Participants were recruited from October 2015 to February 2016 in the Metabolic Unit of the Centre for Nutrition Research of the University of Navarra. Major exclusion criteria included suffering cardiovascular disease, type 1 diabetes, type 2 diabetes treated with insulin, pregnant or lactating women, use of medications that affect body weight, weight change >3kg within 3 months before the start of the intervention, unstable dose of medication for hyperlipidemia, for type 2 diabetes treated with hypoglycemic and/or for hypertension.

The study protocol was approved by the Research Ethics Committee of the University of Navarra (ref. 132/2015). The research was performed in accordance to the ethical guidelines of the Declaration of Helsinki [13]. All participants provided written informed consent after they received an information sheet and additional verbal explanation of the protocol.

2.2. Diet intervention

Energy requirements were individually evaluated from resting energy expenditure, according to the Mifflin formula, multiplied for physical activity level calculated by a short 24h physical activity questionnaire [14–16]. Diets presented the following target macronutrient composition: low-fat diet: 60% of total energy from carbohydrate, 18% from protein and 22% from fat; and moderately-high-protein diet: 40% of total energy from carbohydrate, 30% from protein and 30% from fat. Prescribed diets provided a 30% restriction of the total energy expenditure estimated for each subject. No initial prescribed diets had less than 1200 kcal/day.

Subjects were randomly assigned to one of the two diets by a specific logarithm design for the study by MATLAB using stratified block randomization according to gender, age groups (<45 years and ≥45 years), ethnicity (Caucasian and Hispanic) and BMI (overweight, BMI 25-29.9 kg/m2; and obesity, BMI 30-40 kg/m2).

Compliance analysis to the recommended diet of the participants was conducted taking into account a 3-day-weighed food record (2 weekdays and 1 weekend day) at two times the 8th week and at the end of the intervention period (16th week). Total energy intake and nutrient content were determined using validated Spanish food composition tables and appropriate software [17–19].

217 Chapter 6

2.3. General measurements

Anthropometric and body composition determinants were taken at the beginning and at the end of the study in fasting conditions with the subjects in their underwear by as described elsewhere [20]. Body weight was assessed using the Tanita BC-418 (Tanita, Tokyo, Japan) and height was measured using a wall-mounted stadiometer. BMI was calculated dividing weight (kg) by the square of height (m). Waist circumference (WC) was measured, using a stretchable tape measure, at the midway between the lower margin of the least rib and the top of iliac crest or according to the circumference at the level of the umbilicus if it was not possible to identify the least rib or the iliac crest. Body composition and distribution (fat mass, lean mass, trunk fat, android fat, gynoid fat and visceral fat) was analyzed by dual energy x-ray absorptiometry (DEXA) scan (DEXA Lunar Prodigy, GE Medical Systems, Madison, WI, USA). Baseline dietary intake was analyzed by a previously validated 137 item food frequency questionnaire [21–23].

2.4. Genotyping

For genotyping epithelial buccal cells were collected using a liquid-based kit (ORAcollect-DNA, OCR-100, DNA genotek, Ottawa, Canada). Genomic DNA was extracted with the Maxwell 16 Buccal Swab LEV DNA Purification Kit in the Maxwell 16 instrument (Promega, Madison, USA). ADCY3 rs10182181 was genotyped by Next-Generation Sequencing using a pre-designed SNP panel (Ion AmpliSeq Custom NGS DNA Panels, Thermo Fisher Scientific Inc, Waltham, MA, USA), which was validated in the Ion Torrent PGM (Thermo Fisher Scientific Inc, Waltham, MA, USA). Data were analyzed with the Torrent Variant Caller plugin for the Ion Torrent Sequencing platform and R software.

2.5. Statistical analyses

The primary end points of the present study were changes in anthropometric (weight, waist circumference) and body composition (fat mass, lean mass, trunk fat, android fat gynoid fat and visceral fat). General linear models for continuous variables and the chi-squared test for categorical variables were applied for the comparison according to genotype groups at baseline. Chi-squared test was also used to assess the Hardy-Weinberg equilibrium (HWE). The adherence to the diet across the genotype groups were examined using general linear models adjusted for age and gender. Multivariate general linear models were applied to test changes in primary outcomes according to genotype groups after adjustment for covariates (model 1:

218 Results age, gender and the respective baseline variable; and model 2: model 1 plus BMI at baseline). The interaction term (e.g. ADCY3 genotype x diet) was included in the models in order to test potential gene-diet interactions. The polymorphism was analyzed for additive and co- dominant effects. For the statistical analysis, STATA/SE version 12.0 (StataCorp, Collegue Station, TX, USA) was used. The statistical significant was considered at p < 0.05.

3. RESULTS

Baseline characteristics were similar among participants assigned to the moderately-high- protein diet and the low-fat diet (Table 1). Forty (27%) subjects failed to complete the dietary intervention with no significant differences (p=0.83) among diets (Figure 1). The changes in body fatness and body composition measurements were statistically significant in both diets. The mean weight loss was 7.6 kg in the high-moderately-protein diet, and 8.1 kg in the low-fat diet, with no group difference (p=0.62). There were no differences either in the changes in BMI and WC as well as in any of the assessed body composition measurements (fat mass, lean mass, trunk fat, android fat, gynoid fat and visceral fat) (Table 1).

The minor allele frequency (G allele) of the ADCY3 rs10182181 genetic variant was 0.51 among study participants. The polymorphism was in Hardy-Weinberg Equilibrium (HWE) (p>0.05). Baseline characteristics of the participants included in the present study were also analyzed according to ADCY3 rs10182181 genetic variant (Table 2). The distribution of gender and diet group did not differ depending on the genotype. No significant difference was found in dietary intake and in anthropometric and body composition measurements at baseline examination, whereas a significant difference was observed for age (p=0.007). Those subjects carrying the G allele presented a lower average age than homozygous subjects for the A allele.

In the moderately-high-protein diet group, the targets of macronutrient intakes during the intervention were achieved (Table 3). Whereas the target of macronutrient intake were not fully achieved among subjects of the low-fat diet. However, the reported dietary intake confirmed that participants modified their intakes of macronutrients in the direction of the intervention and significant differences were observed for the intake of fat, protein and carbohydrates between groups (all p values <0.001). There were no significant differences in energy and macronutrient intakes depending on the ADCY3 rs10182181 genotype and group diet (Table 3).

In the present study, there were no significant associations of the ADCY3 polymorphism with changes in anthropometric and body composition measurements during the 16-week dietary

219 Chapter 6 intervention after adjustment for age, sex, the baseline value for the respective outcome, BMI at baseline and dietary group (data not shown). However, we found significant interactions between ADCY3 rs10182181 genotype and dietary intake on changes in body fatness and body composition measurements (Table 4). Specifically, the rs10182181 genetic variant interacted with dietary intake on changes in weight, WC, fat mass, percentage of fat mass, percentage of lean mass, trunk fat, android fat, gynoid fat and visceral fat (model 2 all p for interaction <0.05) (Table 4). After adjusted for age, gender, the respective baseline variable and BMI at baseline the G allele was significantly associated with a lower decrease of fat mass, trunk fat and android fat, and a greater decrease of lean mass in the moderately-high protein diet group (p=0.03, p=0.03, p=0.01 and p=0.04, respectively). Conversely, carrying the G allele was associated with a greater decrease in trunk fat, android fat, gynoid fat and visceral fat when consuming a low-fat diet, once adjusted for covariates (p=0.04, p=0.02, p=0.03 and p=0.02, respectively). Similar trends were found for the co-dominant model (Figure 2).

4. DISCUSSION

The current study reported for the first time a significant gene-diet interaction between ADCY3 rs10182181 genetic variant and dietary macronutrient composition of low calorie diets on changes in anthropometric and body composition measurements. Among individuals with the rs10182181 G-allele consuming the low-fat diet showed greater effect on changes in trunk fat, android fat and gynoid fat, compared with the moderately-high-protein diet over the 16-week dietary intervention.

The ADCY3 gene encodes for an enzyme that converts the ATP in cAMP, which is a second messenger used for intracellular signal transduction [3]. This messenger is involved in a large number of physiological metabolic processes including the regulation of carbohydrate and lipid metabolisms, and the development and function of the adipose tissue regulating the expression of genes involved in adipogenesis, thermogenesis and lipolysis [3,24]. In 2008, for the first time Nordman et al. reported that the ADCY3 rs2033655 and rs1968482 genetic variants were related to obesity but not with type 2 diabetes in a Swedish male population [6]. Subsequently, the authors replicated the genetic association in a large cohort of Chinese adults [7]. In this study, the genetic variants rs1127568, rs7604576 and rs753529 were significantly associated with obesity. Moreover, several GWAS have identified ADCY3 as a gene associated with obesity in adult and children populations [6–12]. For example, the largest meta-analysis of GWAS on BMI carried out-to-date found that the ADCY3 rs10182181 and rs713586 variants

220 Results were associated with BMI [12]. In another GWAS the rs713586 polymorphism, which is in strong linkage disequilibrium with the rs10182181 polymorphism, was identified as an associated BMI variant as well as an expression quantitative trait loci (eQTL) since it was related to ADCY3 gene expression in different tissues (lymphocytes, omental fat and blood) [8]. Interestingly, after an in silico analysis we have confirmed that the rs10182181 polymorphism is in the promoter region of the ADCY3 gene. Particularly, it is located in a binding site for transcription factors including USF1, POLR2A, BHLHE40, JUNB and CREM. This observation suggests that rs10182181 polymorphism may be important for the biological function of the ADCY3 gene.

In accordance with the studies in humans, the Adcy3 knockout mice developed obesity characterized by an increase in fat mass and larger adipocytes [25]. Furthermore, the authors reported that Adcy3-/- mice exhibited reduced physical activity, increased food intake, and leptin insensitivity; and speculated that these phenotypic changes could be associated with disruption of cAMP signaling in primary cilia of the hypothalamus. Recently, the same group using a floxed Adcy3 mouse strain determined that Adcy3 in the hypothalamus regulated energy expenditure [26]. Apart from the hypothalamus, it has been found that Adcy3 is over- expressed in pancreatic islets of non-obese-type 2 diabetic Goto-Kakizaki rats, playing an important role in insulin secretion regulation [5]. There is also some evidence to suggest that Adcy3 may play specific physiological roles in major depression and sleep disruption, which are disorders strongly associated with the obesity phenotype [27]. Moreover, Adcy3 may functionally couple to melanocortin 4 receptor (Mc4r) in the hypothalamus, because activation of adenylyl cyclase activity by alpha-melanocyte stimulating hormone downstream in the leptin pathway is required for the anorectic activity of leptin [28]. In fact, Mc4r and Adcy3 knockout mice exhibit similar phenotypes including obesity and hyperinsulinemia [25,29].

In the present study, macronutrient distribution significantly modified the effect of the ADCY3 genetic variants on changes in anthropometric and body composition measurements. The participants with the G allele of the rs10182181 ADCY3 variant showed a greater decreased in fat mass, trunk fat, android fat, gynoid fat and visceral fat when consuming a low-fat diet. In this sense, it has been reported that a high fat diet decreased the Adcy3 expression in white adipose tissue, liver and muscle [30]. This haploinsufficiency confers decreased expression of genes involved in thermogenesis, fatty acid oxidation and insulin signaling in mice and conversely, it enhanced the expression of genes related to adipogenesis in peripheral tissues. Moreover, mice with a gain-of-function mutation in Adcy3 presented increased Adcy3 activity and cAMP production and consequently the mutation protects mice from high fat diet-induced

221 Chapter 6 metabolic disorders [31]. However, the mechanisms underlying the modulation of macronutrient intake on the ADCY3 genetic variant are not fully understood and further experimental studies are needed.

To our knowledge, this is the first investigation to analyze the effect of the ADCY3 genetic variant on changes in fatness and body composition in response to two weight loss diets with different macronutrient composition. However, several limitations should be considered. First, analyzing the allocated diet (moderately high-protein vs low-fat diet) rather than actual diet may have obscured the interactions found in this study. Nevertheless, using reported intake has its own problems with potential misreporting and would challenge the advantage of the study’s randomized design. Second, it is difficult to determine which macronutrient plays the key role behind the observed interactions because both weight loss diets differed in the content of carbohydrates, protein and fat. Third, the participants of the study were of self- reported European ancestry. Thus, it is unknown whether our results can be generalized to other ethnic groups.

In conclusion, carriers of the minor allele of ADCY3 genotypes might have a better response to a weight-loss dietary intervention by choosing a low-fat diet than a moderately-high-protein diet. Identifying gene-diet interactions on response to metabolic features may assists therapists in assign a more personalized and successful treatments, which could improve long term weight management [32].

Acknowledgments

The authors thank all of the participants for their dedication and contribution to the research. Moreover, we wish to thank Blanca E. Martínez de Morentín, Salomé Pérez, Iosune Zubieta, Laura Olazarán, Ana Lorente as well as Beatriz Ramírez for excellent technical assistance in the University of Navarra and the Center for Applied Medical Research (CIMA).

222 Results

REFERENCES

1. Zhang Y, Liu J, Yao J, Ji G, Qian L, Wang J, et al. Obesity: pathophysiology and intervention. Nutrients. 2014;6: 5153–5183. 2. Mariman ECM, Vink RG, Roumans NJT, Bouwman FG, Stumpel CTRM, Aller EEJG, et al. The cilium: a cellular antenna with an influence on obesity risk. Br J Nutr. 2016;116: 576–592. 3. Wu L, Shen C, Seed Ahmed M, Östenson C-G, Gu HF. Adenylate cyclase 3: a new target for anti-obesity drug development. Obes Rev. 2016;17: 907–914. 4. Abdel-Halim SM, Guenifi A, He B, Yang B, Mustafa M, Höjeberg B, et al. Mutations in the promoter of adenylyl cyclase (AC)-III gene, overexpression of AC-III mRNA, and enhanced cAMP generation in islets from the spontaneously diabetic GK rat model of type 2 diabetes. Diabetes. 1998;47: 498–504. 5. Seed Ahmed M, Kovoor A, Nordman S, Abu Seman N, Gu T, Efendic S, et al. Increased expression of adenylyl cyclase 3 in pancreatic islets and central nervous system of diabetic Goto-Kakizaki rats: a possible regulatory role in glucose homeostasis. Islets. 2012;4: 343–8. 6. Nordman S, Abulaiti A, Hilding A, Långberg E-C, Humphreys K, Ostenson C-G, et al. Genetic variation of the adenylyl cyclase 3 (AC3) locus and its influence on type 2 diabetes and obesity susceptibility in Swedish men. Int J Obes. 2008;32: 407–412. 7. Wang H, Wu M, Zhu W, Shen J, Shi X, Yang J, et al. Evaluation of the association between the AC3 genetic polymorphisms and obesity in a Chinese Han population. PLoS One. 2010;5: e13851. 8. Speliotes EK, Willer CJ, Berndt SI, Monda KL, Thorleifsson G, Jackson AU, et al. Association analyses of 249,796 individuals reveal 18 new loci associated with body mass index. Nat Genet. 2010;42: 937–948. 9. Monda KL, Chen GK, Taylor KC, Palmer C, Edwards TL, Lange LA, et al. A meta-analysis identifies new loci associated with body mass index in individuals of African ancestry. Nat Genet. 2013;45: 690–696. 10. Berndt SI, Gustafsson S, Magi R, Ganna A, Wheeler E, Feitosa MF, et al. Genome-wide meta-analysis identifies 11 new loci for anthropometric traits and provides insights into genetic architecture. Nat Genet. US Department of Health and Human Services, Division of Cancer Epidemiology and Genetics, National Cancer Institute, US National Institutes of Health, Bethesda, Maryland, USA.; 2013;45: 501–512. 11. Graff M, Ngwa JS, Workalemahu T, Homuth G, Schipf S, Teumer A, et al. Genome-wide analysis of BMI in adolescents and young adults reveals additional insight into the effects of genetic loci over the life course. Hum Mol Genet. 2013;22: 3597–3607. 12. Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Day FR, et al. Genetic studies of body mass index yield new insights for obesity biology. Nature. 2015;518: 197–206. 13. World Medical Association. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. United States; 2013;310: 2191–2194. 14. Panel on Macronutrients, Panel on the Definition of Dietary Fiber, Subcommittee on Upper Reference Levels of Nutrients, Sub- committee on Interpretation and Uses of Dietary Reference Intakes and the SC on the SE of DRI. Dietary reference intakes for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein, and amino acids (Macronutrients). Washington; 2005. 15. Mifflin MD, St Jeor ST, Hill L a, Scott BJ, Daugherty S a, Koh YO. A new predictive equation in healthy individuals for resting energy. Am J Clin Nutr. 1990;51: 241–247. 16. Seagle HM, Strain GW, Makris A, Reeves RS. Position of the American Dietetic Association: weight management. J Am Diet Assoc. 2009;109: 330–346. 17. Mataix Verdú J. Tabla de compicion de alimentos. 5th ed. Granada: Universidad de Granada; 2009.

223 Chapter 6

18. Moreiras O, Carbajal Á, Cabrera L, Cuadrado C. Tablas de composición de alimentos. 16th ed. Madrid: Piramide; 2013. 19. Farrán A, Zamora R, Cervera P. Tablas de composición de alimentos del CESNID. 2nd ed. Madrid: Mc Graw Hill; 2009. 20. Perez S, Parra MD, Martínez de Morentin BE, Rodríguez CM, Martínez JA. Evaluación de la variabilidad intraindividual de la medida de composición corporal mediante bioimpedancia en voluntarias sanas y su relación con el índice de masa corporal y el pliegue tricipital. Enfermería Clínica. 2006;15: 343–347. 21. Martin-Moreno JM, Boyle P, Gorgojo L, Maisonneuve P, Fernandez-Rodriguez JC, Salvini S, et al. Development and validation of a food frequency questionnaire in Spain. Int J Epidemiol. 1993;22: 512–9. 22. De la Fuente-Arrillaga C, Vázquez Ruiz Z, Bes-Rastrollo M, Sampson L, Martinez-González MA. Reproducibility of an FFQ validated in Spain. Public Health Nutr. 2010;13: 1364–1372. 23. Fernández-Ballart JD, Piñol JL, Zazpe I, Corella D, Carrasco P, Toledo E, et al. Relative validity of a semi- quantitative food-frequency questionnaire in an elderly Mediterranean population of Spain. Br J Nutr. 2010;103: 1808–16. 24. Rogne M, Taskén K. Compartmentalization of cAMP signaling in adipogenesis, lipogenesis, and lipolysis. Horm Metab Res. 2014;46: 833–840. 25. Wang Z, Li V, Chan GCK, Phan T, Nudelman AS, Xia Z, et al. Adult type 3 adenylyl cyclase-deficient mice are obese. PLoS One. 2009;4: e6979. 26. Cao H, Chen X, Yang Y, Storm DR. Disruption of type 3 adenylyl cyclase expression in the hypothalamus leads to obesity. Integr Obes Diabetes. 2016;28: 1304–1314. 27. Chen X, Luo J, Leng Y, Yang Y, Zweifel LS, Palmiter RD, et al. Ablation of type III adenylyl cyclase in mice causes reduced neuronal activity, altered sleep pattern, and depression-like phenotypes. Biol Psychiatry. Elsevier; 2016;80: 836–848. 28. Qiu L, LeBel RP, Storm DR, Chen X. Type 3 adenylyl cyclase: A key enzyme mediating the cAMP signaling in neuronal cilia. Int J Physiol Pathophysiol Pharmacol. 2016;8: 95–108. 29. You P, Hu H, Chen Y, Zhao Y, Yang Y, Wang T, et al. Effects of melanocortin 3 and 4 receptor deficiency on energy homeostasis in rats. Sci Rep. 2016;6: 34938. 30. Tong T, Shen Y, Lee H-W, Yu R, Park T. Adenylyl cyclase 3 haploinsufficiency confers susceptibility to diet- induced obesity and insulin resistance in mice. Sci Rep. 2016;6: 34179. 31. Pitman JL, Wheeler MC, Lloyd DJ, Walker JR, Glynne RJ, Gekakis N. A gain-of-function mutation in adenylate cyclase 3 protects mice from diet-induced obesity. PLoS One. 2014;9: e110226. 32. Goni L, Cuervo M, Milagro FI, Mart JA. Future Perspectives of Personalized Weight Loss Interventions Based on Nutrigenetic, Epigenetic, and Metagenomic Data. J Nutr. 2016;146: 905–912.

224 Results

Table 1. Baseline characteristics and after 16-weeks of intervention concerning anthropometric and body composition measurements

Moderately-high-protein Low-fat a b Baseline Change Baseline Change p p (n=72) (n=53) (n=75) (n=54) Age (years) 45.6 (11.2) - 47.5 (8.6) - 0.28 - Female sex 48 (66.7) - 51 (68.0) - 0.87 - Body weight (kg) c 88.5 (13.2) -7.6 (4.0) 88.8 (13.3) -8.1 (4.1) 0.76 0.62 BMI (kg/m2) c 31.7 (3.6) -2.7 (1.4) 32.1 (3.8) -2.9 (1.3) 0.56 0.71 WC (cm) c 103.5 (11.2) -8.4 (4.4) 103.8 (10.6) -8.8 (4.5) 0.99 0.73 Body composition Fat mass (%) c 42.6 (9.1) -4.8 (5.3) 41.9 (6.9) -3.9 (2.5) 0.37 0.13 Lean mass (%) c 54.2 (9.2) 4.6 (5.4) 55.0 (6.6) 3.6 (2.4) 0.34 0.12 Trunk fat (kg) c 20.3 (5.0) -4.1 (2.5) 20.6 (4.7) -4.3 (2.4) 0.81 0.87 Android fat (kg) c 3.6 (1.0) -0.8 (0.5) 3.6 (0.9) -0.9 (0.5) 0.78 0.98 Gynoid fat (kg) c 5.9 (1.5) -0.9 (0.7) 6.0 (1.7) -1.0 (0.5) 0.74 0.84 Visceral fat (kg) c 1.5 80.9) -0.5 (0.4) 1.7 (0.9) -0.5 (0.4) 0.37 0.92 BMI, Body mass index, WC, Waist circumference Data are expressed as mean (SD) or n (%) a Comparison of baseline characteristic by dietary group b Comparison of changes in anthropometric and body composition measurements by dietary group c Adjusted for age and gender

225 Chapter 6

Table 2. Baseline characteristics concerning dietary and body composition measurements of the participants depending on the ADCY3 rs10182181 genetic variant

AA AG GG p (n=35) (n=74) (n=38) Age (years) 50.9 (9.0) 46.0 (9.5) 43.8 (10.7) 0.007 Sex 0.16 Male 11 (31.4) 20 (27.0) 17 (44.7) Female 24 (68.6) 54 (73.0) 21 (55.3) Diet group 0.84 Moderately high protein 16 (45.7) 36 (48.7) 20 (52.6) Low fat 19 (54.3) 38 (51.3) 18 (47.4) Dietary intake per day Energy (kcal) 2907 (849) 3023 (855) 3038 (1001) 0.78 Protein (%) 16.7 (3.4) 16.5 (2.5) 17.1 (3.1) 0.55 Fat (%) 38.8 (6.2) 39.6 (6.0) 40.0 (5.2) 0.67 Carbohydrate (%) 42.8 (8.5) 41.9 (6.5) 40.8 (7.0) 0.52 Body weight (kg) 88.2 (13.1) 87.6 (12.6) 91.0 (14.6) 0.43 BMI (kg/m2) 32.7 (3.9) 31.6 (3.5) 31.9 (3.6) 0.34 WC (cm) 105.9 (10.2) 102.2 (10.5) 104.5 (12.0) 0.22 Body composition a Fat mass (%) 42.1 (5.4) 42.2 (8.9) 42.4 (8.5) 0.99 Lean mass (%) 54.8 (5.1) 54.7 (8.7) 54.4 (8.5) 0.98 Trunk fat (kg) 20.7 (3.9) 19.9 (4.8) 21.2 (5.7) 0.40 Android fat (kg) 3.6 (0.8) 3.5 (0.9) 3.8 (1.2) 0.30 Gynoid fat (kg) 5.8 (1.5) 6.0 (1.7) 6.0 (1.6) 0.87 Visceral fat (kg) 1.7 (0.9) 1.5 (0.9) 1.7 (1.0) 0.36 BMI, Body mass index; WC, Waist circumference Data are expressed as mean (SD) or n (%) a Data available for 146 participants (AA=34, AG=74, GG=38)

226 Results

Table 3. Dietary intake by the ADCY3 rs10182181 genetic variant and diet group during the intervention

All population AA AG GG p Moderately-high

protein diet a Energy (kcal) 1340 (286) 1297 (180) 1294 (238) 1490 (414) 0.11 Protein (%) 28.9 (4.3) 29.2 (3.5) 29.7 (4.2) 26.7 (4.4) 0.12 Fat (%) 30.8 (5.0) 31.7 (5.4) 30.1 (4.9) 32.1 (5.1) 0.43 Carbohydrate (%) 42.5 (4.3) 42.1 (3.7) 41.9 (3.8) 44.2 (5.7) 0.27 Low-fat diet b Energy (kcal) 1324 (236) c 1313 (274) 1291 (231) 1429 (185) 0.29 Protein (%) 22.3 (3.7) d 22.0 (4.2) 22.2 (3.2) 22.8 (4.8) 0.89 Fat (%) 27.0 (5.7) d 25.6 (8.3) 27.4 (3.7) 27.6 (6.6) 0.61 Carbohydrate (%) 53.3 (7.1) d 54.5 (8.5) 53.1 (6.5) 52.1 (7.5) 0.72 a Data were available for 52 individuals (AA n=10, AG n=30, GG n=12) b Data were available for 51 individuals (AA n=13, AG n=28, GG n=10) c p=0.76 when it was compared with moderately-high-protein diet d p<0.001 when it was compared with moderately-high-protein diet

227 Chapter 6

Table 4. Effect of the ADCY3 rs10182181 genetic variant on changes in body fatness and composition in response to moderately-high-protein/low-fat diet at 16 weeks of diet intervention (additive model)

Moderately-high-protein Low-fat p interaction b (SE) p b (SE) p Model 1 D Weight (kg) 0.95 (0.82) 0.25 -1.32 (0.69) 0.06 0.02 D WC (cm) 1.20 (0.95) 0.21 -0.88 (0.84) 0.30 0.06 D Fat mass (kg) 1.26 (0.76) 0.10 -0.95 (0.57) 0.10 0.01 D Fat mass (%) 1.65 (0.92) 0.08 -0.54 (0.44) 0.22 0.03 D Lean mass (kg) -0.54 (0.96) 0.57 -0.15 (0.30) 0.60 0.58 D Lean mass (%) -1.61 (0.90) 0.08 0.48 (0.42) 0.25 0.04 D Trunk fat (kg) 0.89 (0.50) 0.08 -0.72 (0.41) 0.08 0.008 D Android fat (kg) 0.22 (0.11) 0.049 -0.18 (0.08) 0.02 0.002 D Gynoid fat (kg) 0.18 (0.14) 0.20 -0.19 (0.08) 0.03 0.02 D Visceral fat (kg) 0.08 (0.06) 0.16 -0.12 (0.05) 0.02 0.005 Model 2 D WC (cm) 1.72 (0.94) 0.07 -0.61 (0.81) 0.45 0.03 D Fat mass (kg) 1.49 (0.78) 0.06 -1.08 (0.57 0.06 0.006 D Fat mass (%) 2.21 (0.89) 0.01 -0.57 (0.43) 0.20 0.01 D Lean mass (kg) -0.41 (0.92) 0.65 -0.15 (0.30) 0.63 0.52 D Lean mass (%) -2.13 (0.84) 0.01 0.50 (0.41) 0.23 0.01 D Trunk fat (kg) 1.15 (0.52) 0.03 -0.85 (0.40) 0.04 0.006 D Android fat (kg) 0.28 (0.11) 0.01 -0.19 (0.08) 0.02 0.001 D Gynoid fat (kg) 0.27 (0.14) 0.05 -0.20 (0.09) 0.03 0.008 D Visceral fat (kg) 0.10 (0.06) 0.09 -0.11 (0.05) 0.02 0.005 WC, Waist circumference β represents changes in outcomes for the increasing number of G allele of the ADCY3 rs10182181 variant Model 1: Adjusted for age, gender and the respective baseline variable Model 2: Model 1 plus BMI at baseline

228 Results

Figure 1. Flow-chart of the participants

229 Chapter 6

Moderately high protein Low fat Moderately high protein Low fat 0 0

-2 -2

-4 AA -4 AA

AG AG

Weight (kg) -6 -6

∆ GG GG -8 -8 as icmeec (cm) ∆ Waist circumference -10 -10

p=0.52 p=0.03 p=0.36 p=0.35 -12 p interaction=0.04 -12 p interaction=0.12

Moderately high protein High fat 8 p interaction=0.06 0 p=0.06 p=0.47 7

-2 6

5 AA -4 AA 4 AG AG

∆ Fat mass (%) GG -6 3 GG ∆ Lean mass (%) 2 -8 1

p=0.07 p=0.43 0 -10 p interaction=0.06 Moderately high protein Low fat

Moderately high protein Low fat Moderately high protein Low fat 0

-1 -0.2

-0.4 AA AA -3 AG -0.6 AG GG -0.8 GG ∆ Trunk fat (kg) -5 nri fat (kg) ∆ Android -1

-1.2 p=0.24 p=0.09 -7 p interaction=0.04 p=0.12 p=0.048 -1.4 p interaction=0.009

Moderately high protein Low fat 0 Moderately high protein Low fat 0 -0.2 -0.1 -0.4 -0.2 AA -0.6 AA AG -0.3 -0.8 AG GG -1 -0.4 GG ∆ Gynoid fat (kg) ∆ Visceral fat (kg) -1.2 -0.5

-1.4 -0.6 p=0.31 p=0.07 -1.6 p interaction=0.07 p=0.18 p=0.04 -0.7 p interaction=0.02

Figure 2. Effect of the ADCY3 rs10182181 genetic variant on changes in body fatness and composition in response to moderately-high-protein/low-fat diet at 16 weeks of diet intervention (co-dominant model)

230

CHAPTER 7

Gene-gene interplay and gene-diet interactions involving the MTNR1B rs10830963 variant with body weight loss

Goni L.1,2, Cuervo M.1,2,3,4, Milagro F.I.1,2,3,4, Martínez J.A.1,2,3,4

1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Navarra, Spain 2 Centre for Nutrition Research, University of Navarra, Navarra, Spain 3 Navarra Institute for Health Research (IdiSNA), Navarra, Spain 4 Biomedical Research Centre Network in Physiopathology of Obesity and Nutrition (CIBERobn), Institute of Health Carlos III, Madrid, Spain

J Nutrigenet Nutrigenomics 2014 DOI: 10.1159/000380951 Impact factor (2014): 2.000 48/77 Nutrition & Dietetics, Q3 115/167 Genetics & Heredity, Q3

Results

233 Chapter 7

ABSTRACT

Background/Aims: Investigation of the genetic make-up may facilitate the implementation of more personalized nutritional interventions. The aims were to examine whether the rs10830963 MTNR1B polymorphism affects weight loss in response to a hypocaloric diet and to find potential gene-gene interplays and gene-diet interactions.

Methods: 167 subjects enrolled in a personalized nutritional intervention for weight loss (3-6 weeks), were examined for anthropometric measurements, dietary habits and physical activity at baseline and at a first follow-up visit. Three polymorphisms, which have been previously associated with body weight regulation, rs10830963 (MTNR1B), rs9939609 (FTO) and rs17782313 (MC4R), were analyzed using Luminex® 100/200TM System.

Results: After adjusted for covariates, females with the rs10830963 CG/GG genotype showed lower weight loss than those with the CC genotype. In the total population, carriers of variant alleles of both, FTO and MC4R, showed a significant association with MTNR1B and weight loss outcome. Moreover, among women, higher total protein and animal protein intakes were associated with a lower weight loss in G allele carriers of the MTNR1B variant.

Conclusions: Our data evidenced that rs10830963 MTNR1B polymorphism could be associated with individual differences in weight loss induced by a hypocaloric diet. This association was influenced by FTO and MC4R loci and modified by baseline protein intake.

Key words

MTNR1B, FTO, MC4R, obesity, body weight loss, gene-gene interplay, gene-diet interaction

234 Results

1. INTRODUCTION

Obesity has reached epidemic proportions becoming a major global health challenge [1]. Therefore, a number of strategies have been investigated in order to induce a negative energy balance and body weight loss, such as a reduction of energy intake, an increase in physical activity, behavioral approaches and pharmacological or surgical treatments [2, 3]. However, individual responses to body weight loss interventions vary widely and several studies have aimed to identify psychological, behavioral and personal predictors of this variability [4-6]. In this context, genetic factors have been described to be associated with adiposity and body weight control, since there are diverse genes involved in the regulation of energy expenditure, appetite, thermogenesis, adipogenesis, insulin resistance and lipid metabolism [7-9].

Circadian disruptions may contribute to obesity and metabolic related traits (glucose intolerance, type 2 diabetes mellitus, hypertension, dyslipidemia or cardiovascular disease) [10]. In fact, recent studies on the etiology of obesity have evaluated the role of common polymorphisms located in genes involved in circadian systems [11]. Among them, melatonin receptor 1B (MTNR1B) might be implicated in body weight regulation, given that this gene encodes a receptor for melatonin, a hormone involved in energy balance and body weight status [12, 13].

Some studies have reported that two polymorphisms, rs10830963 and rs8192552 located in the MTNR1B gene are associated with obesity [14, 15]. However, to date, apparently the existence of a relationship between rs10830963 and changes in body weight or other anthropometric measurements induced by a weight loss intervention need to be elucidated [16, 17]. In this regard, a recent study has demonstrated an association between rs10830963 MTNR1B polymorphism and changes in respiratory quotient (RQ) after a body weight loss program, suggesting a relationship with energy expenditure [17]. Interestingly, it has been demonstrated that MTNRB1 is associated with type 2 diabetes mellitus and more recently with gestational diabetes mellitus [18, 19]. The effect of the MTNR1B gene on the obesity phenotype is still controversial, but its interactions with other polymorphisms near or within genes involved in body weight loss, such as fat mass and the obesity-associated gene (FTO) and melanocortin-4 receptor gene (MC4R), and gene-diet interactions may partly explain the discrepancies [20, 21].

The aims of this study were to investigate the influence of the rs10830963 MTNR1B variant in relation to body weight loss in response to a hypocaloric diet intervention and to examine potential gene-gene interplays and gene-diet interactions.

235 Chapter 7

2. MATERIALS AND METHODS

2.1. Subjects

The current study enrolled men and women of Caucasian ancestry, who voluntarily attended community pharmacies in 7 cities of Spain (Barcelona, Zaragoza, La Coruña, Pontevedra, Madrid, Granada and Málaga) [22]. A total of 465 individuals complied with the first follow-up visit. Participants were excluded for the following reasons: body mass index (BMI) <25 kg/m2 (n=91), time between baseline visit and first follow-up visit under 3 weeks or above 6 weeks (n=204) and missing values for anthropometric measurements (n=3). Therefore, the screened group encompassed 167 subjects.

Individuals were specifically asked if they would be willing to take part anonymously in the research study. Only those who provided written informed consent for participation were enrolled. The survey was performed in accordance with the principles of the Declaration of Helsinki and patient data were codified to guarantee anonymity. The Research Ethics Committee of the University of Navarra gave confirmation of fulfillment of the ethical standards affecting this research (Ref. 2410/2014).

2.2. Anthropometric measurements, physical activity and usual dietary intake

Anthropometric measurements at baseline and at the first follow-up visit were collected by trained nutritionists using a standardized protocol. Weight and height were measured, with subjects wearing light clothes and no shoes, using a digital scale (Tanita BF-522W, Tanita Corporation, Tokyo, Japan) and a portable stadiometer (Leicester Tanita), respectively. Body fat mass (BFM) was analyzed by bioelectrical impedance applying the Tanita BF-522W. Waist and hip circumferences were measured with a flexible and inextensible tape device. Waist circumference was determined at the midway between the lower margin of the last rib and the top of iliac crest or according to the circumference at the level of the umbilicus if it was not possible to identify the last rib or the iliac crest. Meanwhile, hip circumference was measured as the widest circumference over the greater buttocks. BMI, as weight (kg)/ height (m)2, waist- to-hip ratio (waist/hip) and waist-to-height ratio (waist/height) were then calculated.

Physical activity was estimated by a short 24-hour physical activity questionnaire [23]. Individuals were asked to report the number of hours resting and practicing activities (according to intensity of effort: sedentary, low active, active and very active) at work and at leisure time during a week and a weekend day. Next, individual daily physical activity level was

236 Results computed multiplying the average time spent on each group of activities and multiples of physical activity levels [23].

Baseline dietary intake was assessed by a food frequency questionnaire in which staple foods were classified into 19 food groups according to a food exchange list system: whole dairy products, half-fat dairy products and fat-free dairy products, eggs, fat meat and sausages, lean meat, white fish and sellfish, blue fish, vegetables, fruits, nuts, legumes, olive oil, other fats and oils, refined grains, whole grains, pastries and confectionary industry, sugars, water and alcohol drinks. Therefore, food portions on each group have about the same amount of calories and macronutrients. Individuals were asked how often (daily, weekly, monthly or never) they had consumed a choice of each food group during the previous year. Food intakes were converted to energy and nutrient intake data using Spanish food composition tables [24, 25].

2.3. Characteristics of the nutritional intervention

Individual energy requirements were evaluated by computing resting energy expenditure according to the Harris-Benedict formula, and total energy expenditure according to the physical activity level, previously explained [23]. Diets were designed to provide about 600 kcal/day less than the individually estimated total energy expenditure (restriction of 603 ± 15 kcal). The target macronutrient composition of the diets was: 20% of total energy from proteins, 30% from lipids and 50% from carbohydrates. In order to provide personalized nutrition based on both the phenotype and the genotype, nutritionists instructed participants individually about the type of foods they could eat based on a food exchange list system.

2.4. Genotyping

Samples from oral epithelial cells were collected (ORAcollect DNA®, DNAGenotek, Kanata, Canada) in order to extract genomic DNA by QIAcube using QiAmp DNA Mini QIAcube Kit (Qiagen, Hilden, Germany), following the manufacturer’s procedures. Three polymorphisms, rs10830963 of MTNR1B (Chromosome 11, Position 92,708,710), rs9939609 of FTO (, Position 53,820,527) and 17782313 near to MC4R (Chromosome 18, Position 57,851,097) were genotyped. GeneAmp® PCR System 9700 thermal cycler (Applied Biosystems, Foster City, USA) was used according to standardized laboratory protocols to carry out the polymerase chain reactions. Polymerase chain reactions products were hybridized onto oligonucleotide probes attached to microspheres and labeled with streptavidin-

237 Chapter 7 conjugated physoerythrin (MagPlex-TAG Microspheres). These beads were analyzed with Luminex® 100/200TM System (Luminex Corporation, Austin, Texas) [26]. This method is based on the principles of xMAP® Technology, which encompasses polystyrene microspheres internally dyed with various ratios of spectrally distinct fluorochromes, that are detected by a flow cytometry-based instrument [26].

2.5. Statistical analysis

Examination of Hardy-Weinberg equilibrium was carried out by χ2 test. Due to the limited frequency of the variant allele, homozygotes of the minor allele (GG) and heterozygotes (CG) were grouped and compared with major allele homozygotes (CC). Differences in quantitative baseline characteristics among genotype groups (CC vs CG/GG) were analyzed using general linear models of analysis of covariance (ANCOVA) with age, gender, energy intake and physical activity level as covariates. Qualitative variables were analyzed by χ2 test. Change (Δ) in body weight (kg), ∆BMI (kg/m2) and ∆BFM (%) were computed by subtracting the measurement recorded at baseline visit from the measurement at the first follow-up visit. Furthermore, ∆%, which expresses the relative changes of a parameter to baseline, was calculated as ∆/(measurement recorded at baseline visit x 100). In order to assess the association between the rs10830963 MTNR1B polymorphism and changes in body composition measurements, linear regression analyses were adjusted for gender, age, energy restriction, time between visits, FTO, MC4R, and baseline value of each anthropometric variable (body weight, BMI or BFM). The influence of well-known obesity loci on the association between MTNR1B genetic variant and body weight loss was examined by ANCOVA analyses. Possible interactions between the rs10830963 polymorphism and baseline dietary intake on body weight loss were investigated with the likelihood ratio test (adjusted as described above). All statistical procedures were conducted using STATA/SE version 12.0 (StataCorp, College Station, TX, USA). Tests were considered statistically significant at p value <0.05.

3. RESULTS

The minor allele frequency was 0.29, 0.45 and 0.22 for rs10830963 (MTNR1B), rs99396909 (FTO) and rs17782313 (MC4R), respectively. The distributions of the 3 polymorphisms were in Hardy-Weinberg equilibrium (p>0.05). Baseline characteristics of the study sample were categorized by MTNR1B rs10830963 genotype, CC vs CG/GG (table 1). Of the 167 subjects, 75.4% were females. There were no statistically significant differences associated with

238 Results genotype for anthropometric variables and other parameters of interest (after adjusting for gender, age, energy intake and physical activity), and neither were differences found when the sample was stratified by gender. In general, according to the Spanish dietary recommendations, the dietary pattern of the subjects was characterized by a high consumption of fat and a low carbohydrate intake. Regarding energy restriction (603 ± 15 kcal/day) and time between the baseline visit and the first follow-up visit (27 ± 0.5 days), values were similar between CC and CG/GG genotypes.

Linear regression analyses showed the impact of rs10830963 MTNRB1 genotypes on body weight and BMI changes (Table 2). When the total population was taken into account, no association was found between the polymorphism and changes in body composition measurements. By contrast, upon stratifying data by gender, women with the minor G allele of rs10830963 revealed a lower weight loss (B coefficient -0.58, p 0.036) and therefore a lower BMI loss (B coefficient -0.21, p 0.027), after adjusting for age, energy restriction, time between visits, FTO, MC4R and weight or BMI at baseline. However, no association was observed in men. Similar results were found regarding the percentage of ∆ in body weight and BMI after adjusting for confounder variables (data not shown).

The association of the rs10830963 MTNRB1 genetic variant and changes in body weight and BMI was also analyzed in two groups defined by the presence of obesity-related loci (non-risk allele carriers or risk allele carriers of either FTO or MC4R, and risk allele carriers of both FTO and MC4R). In the non-risk allele carriers or risk allele carriers of either FTO or MC4R groups, there was no significant association of the MTNR1B polymorphism with body weight or BMI loss. However, in the risk allele carriers of both FTO and MC4R group, rs10830963 was significantly associated with body weight and BMI loss in the total population (fig. 1) and a trend toward significance (p=0.073) was observed in women (data not shown).

Among women, animal protein intake modified the effect of rs10830963 on the body weight loss induced by the nutritional intervention, after adjusting for age, energy restriction, time between visits, FTO, MC4R and body weight at baseline (fig. 2). Higher total protein and animal protein intakes were associated with lower body weight loss among G allele carrier group. No significant interactions were found between MTNR1B and energy intake or vegetable protein intake in relation to body weight loss.

239 Chapter 7

4. DISCUSSION

The present research shows a possible effect of the rs10830963 polymorphism, located in MTNR1B, in regulating body weight loss, and the influence of obesity-related polymorphisms, such as FTO and MC4R. This is apparently the first study analyzing whether usual dietary intake can modulate the association between a genetic variant located in MTNR1B and body weight loss in response to a nutritional intervention.

Although two previous studies have reported associations of MTNR1B with obesity phenotype in three different European populations, our research failed to confirm such relationship [14, 15]. In the study by Stančáková et al. [14], rs10830963 was significantly associated with BMI among Finnish men, while Andersson et al. [15] detected that other MTNR1B genetic variant (rs8192552) was related, not only with obesity and BMI, but also with waist circumference in two populations of Danish and French origin [14, 15]. Nevertheless, when the association between this polymorphism and waist circumference was adjusted for BMI, no significant relationship was found, indicating that MTNR1B is apparently related to overall obesity rather than to central obesity. On the other hand, an observation similar to our results was reported in a study which looked at overweight and obese in Germany children and adolescents [27]. Thus, controversial and inconclusive results have been found for the association between MTNR1B and the obesity phenotype.

Interestingly, the present study revealed a relationship between a MTNR1B genetic variant and body weight loss associated by energy restriction among women. G allele carriers showed less body weight and BMI loss than CC genotype subjects. In this sense, it should be mentioned that when the analysis considered three genotype groups there were statistically significant differences between CC and CG genotypes, but not between CC and GG genotypes. The relative small sample size of GG genotype group might limit the ability to detect genetic effects. MTNR1B encodes a receptor for melatonin expressed in the retina, the suprachiasmatic nucleus, the circadian rhythm control center, human pancreatic islets and pancreatic β-cells [12, 28]. The anti-obesogenic effect of melatonin is, in part, a result of its role on an adequate energy balance mainly by regulating energy flow to and from the stores, and regulating the energy expenditure through the activation of brown adipose tissue [29]. Additionally, the reduction in melatonin production involves metabolic disturbances (glucose intolerance, central and peripheral insulin resistance or dyslipidemia) and also chronodisruption (circadian impairment on muscle metabolism, of white adipose tissue metabolism, of hepatic metabolism and of insulin synthesis, secretion and action) related to obesity [13]. In contrast to our results, Peter et al. (2012) and Mirzae et al. (2014) found no

240 Results association between rs10830963 and body composition changes after an intensive lifestyle intervention and body weight loss program, respectively [16, 17]. Of particular interest is the reported relationship between the presence of the G allele of the rs10830963 polymorphism and the RQ, previously identified by Mirzae et al. [17]. G allele carriers showed a greater increase in RQ, suggesting a functional role of the MTNR1B in energy expenditure. Therefore, it is possible that rs10830963 may be involved in the regulation of MTNR1B gene expression or the expression of other genes that may influence the role of melatonin in energy balance. In this sense, a previous study reported an increase of the MTNR1B mRNA expression in risk genotype carriers of the rs10830963 in human pancreatic islets, also showing a negative correlation between MTNR1B mRNA levels and insulin secretion [30, 31]. However, rs10830963 MTNR1B genetic variant is located in an intronic region, so additional analyses are required in order to provide clearer evidence on its functional role regarding body weight control.

An apparent paradox is that MTNR1B is not related with obesity in our sample population, while is associated with body weight reduction induced by a hypocaloric diet, most probably because of insufficient power to detect modest effects of SNPs located in this gene. Another possible explanation for this dual finding is that body weight loss is not the reverse of primary weight gain, and both processes imply different mechanisms [32].

The association between rs10830963 polymorphism and body weight loss was exhibited in women but not in men. This finding could be due in part to the small sample size of the group of men. On the other hand, the sex could modify the effects of obesity genes on body composition, due to a complex interplay of genetic and hormonal factors [33]. Thus, Demple et al. [34] detected that MC4R polymorphisms had larger effects on obesity in females than in males. Moreover, previous genome-wide association studies and meta-analyses have demonstrated several anthropometric trait loci significantly associated with body fat distribution measurements in women but not in men [35-37]. Although previous studies have demonstrated sex-specific polymorphisms related with obesity phenotype, more investigations are needed regarding body weight loss depending on gender [38].

It should be noted that, when the linear regression models were not adjusted for FTO and MC4R genotypes, no significant associations were found between the rs10830963 MTNR1B genetic variant and changes in body weight or BMI (B coefficient -0.48, p=0.070 and -0.20, p=0.050; respectively). For this reason, we investigated the association of rs10830963 and changes in body weight and BMI in 2 groups defined by the presence of FTO and MC4R (non- risk allele carriers or risk allele carriers of either FTO or MC4R, and risk allele carriers of both

241 Chapter 7

FTO and MC4R). FTO encodes an enzyme that has been identified as an mRNA demethylase [39]. The mechanism underlying the association of the rs9939609 FTO variant with obesity remains still unclear. In fact, a study found a relationship between rs9939609 and FTO expression whereas another one did not observe such an association [40, 41]. Recently, it has been demonstrated that FTO is functionally connected with the regulation of Iroquois homeobox 3 (IRX3) expression, which is involved in body composition [42]. On the other hand, MC4R encodes one of the key peptide of the melanocortin pathway, which is a hypothalamic regulator of energy balance [43]. As other authors have suggested, the influence of rs17782313 on obesity may be mediated through effects on MC4R expression, as a consequence of its location [44]. Although, there is some ambiguity about the effect of rs9939609 and rs17782313 on body weight loss, they are two of the three polymorphisms most strongly associated with the obesity phenotype [45]. In this sense, we did not found that polymorphisms located in these genes were associated with changes in body weight and BMI, in accordance with other studies [46, 47]. Interestingly, this analysis highlighted that polymorphisms located in FTO and MC4R, at the same time influence the relationship between MTNR1B and body weight loss, maybe due to a synergistic effect. This finding may partly explain the discrepancies about the effect of FTO, MC4R and MTNR1B and obesity phenotype among different studies and populations.

On the other hand other trials have demonstrated that the response of subjects to diet might be different depending on the genotype [48, 49]. In the recruited population, baseline total protein and animal protein intake significantly modified the effect of MTNR1B genetic variant on body weight loss. There is evidence that a high-protein diet affects rhythmic expression of circadian clock genes in mouse peripheral tissues [50]. It should be highlighted that the study by Mirzaei et al. [17] reported a significant interaction between rs10830963 and dietary fat on RQ, but not with protein intake. In contrast, the baseline protein intake has been previously included among pretreatment predictors of body weight loss [51]. In this sense, the thermogenic effect of the dietary protein is well known [52]. We hypothesized that those individuals with a high-protein intake at baseline had adapted to the increased thermogenic response of a high-protein diet and this factor could inhibit the body weight loss. However, the process by which usual protein intake influences body weight loss depending on MTNR1B genetic variant is unknown.

We acknowledge some limitations of this research. The relative small sample size might limit the power to detect the effect of MTNR1B across obesity phenotype and gene-diet interactions. The short time of follow-up may have restricted the ability to detect meaningful

242 Results long-term differences. In this sense, a recent study has detected that the initial body weight loss (at 1 month) is a positive predictor of weight loss after 12 months of intervention [53]. Although the findings of this investigation are based on a population of a nutritional service based on the genetic make-up, the participants included in the study were not necessarily representative of the general population. We did not control, during the intervention period, the intake of some nutrients or foods which could have been involved in body weight reduction. However, due to the short time between the baseline and the first follow-up visit, we assumed that the body weight loss was due to the caloric restriction. As a final point, the time between baseline and the first follow-up visit and also the energy restriction differed between participants. Nevertheless, there were no statistical differences between genotype groups (CC vs CG/GG) and all of the statistical analyses were adjusted for these two variables, time between baseline and first follow-up visit, and energy restriction.

In conclusion, our finding concerning the association of rs10830963 MTNR1B genetic variant with body weight loss induced by a hypocaloric diet provides additional evidence on the importance of the genes involved in the circadian system as risk genes for obesity phenotype, but also their modulation by other polymorphisms involved in obesity phenotype and dietary habits such as protein intake. The investigation of genetic factors and gene-diet interactions may improve the body weight loss interventions because of the identification of vulnerable individuals that will benefit from more personalized dietary recommendations.

Acknowledgements

We thank all the seekers of the Nutrigenetic Service who voluntarily offered their data, and the 5 nutritionists for data collection, as well as Amaia Ibáñez for excellent technical assistance. The authors thank the Linea Especial (University of Navarra; LE/97) and to the Spanish Ministry of Economy and Competitiveness (AGL2013-45554-R project) for financial support, CIBERobn/RETICS schedules (Instituto Carlos III, Madrid, Spain) for assistance in this study and the support from CINFA (Olloki, Spain) concerning the genetic tools and general logistic. The pre-doctoral research grant to Leticia Goñi from the Asociación de Amigos Universidad de Navarra is gratefully acknowledged.

243 Chapter 7

REFERENCES

1. Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono C, et al. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980-2013: A systematic analysis for the global burden of disease study 2013. Lancet. 2014;384:766-81. 2. Kushner RF. Weight loss strategies for treatment of obesity. Prog Cardiovasc Dis. 2014;56:465-72. 3. Abete I, Astrup A, Martinez JA, Thorsdottir I, Zulet MA. Obesity and the metabolic syndrome: Role of different dietary macronutrient distribution patterns and specific nutritional components on weight loss and maintenance. Nutr Rev. 2010;68:214-31. 4. Perez-Cornago A, Ramirez MJ, Zulet MA, Martinez JA. Effect of dietary restriction on peripheral monoamines and anxiety symptoms in obese subjects with metabolic syndrome. Psychoneuroendocrinology. 2014;47:98-106. 5. Burmeister JM, Carels RA. Television use and binge eating in adults seeking weight loss treatment. Eat Behav. 2014;15:83-6. 6. Mariman EC. Human biology of weight maintenance after weight loss. J Nutrigenet Nutrigenomics. 2012;5:13-25. 7. Moreno-Aliaga MJ, Santos JL, Marti A, Martinez JA. Does weight loss prognosis depend on genetic make-up? Obes Rev. 2005;6:155-68. 8. Deram S, Villares SM. Genetic variants influencing effectiveness of weight loss strategies. Arq Bras Endocrinol Metabol. 2009;53:129-38. 9. Goni L, Milagro FI, Cuervo M, Martinez JA. Single-nucleotide polymorphisms and DNA methylation markers associated with central obesity and regulation of body weight. Nutr Rev. 2014;72:673-90. 10. Gomez-Abellan P, Madrid JA, Ordovas JM, Garaulet M. Chronobiological aspects of obesity and metabolic syndrome. Endocrinol Nutr. 2012;59:50-61. 11. Summa KC, Turek FW. Chronobiology and obesity: Interactions between circadian rhythms and energy regulation. Adv Nutr. 2014;5:312S-9S. 12. Barrenetxe J, Delagrange P, Martinez JA. Physiological and metabolic functions of melatonin. J Physiol Biochem. 2004;60:61-72. 13. Cipolla-Neto J, Amaral FG, Afeche SC, Tan DX, Reiter RJ. Melatonin, energy metabolism, and obesity: A review. J Pineal Res. 2014;56:371-81. 14. Stancakova A, Kuulasmaa T, Paananen J, Jackson AU, Bonnycastle LL, Collins FS, Boehnke M, Kuusisto J, Laakso M. Association of 18 confirmed susceptibility loci for type 2 diabetes with indices of insulin release, proinsulin conversion, and insulin sensitivity in 5,327 nondiabetic Finnish men. Diabetes. 2009;58:2129-36. 15. Andersson EA, Holst B, Sparso T, Grarup N, Banasik K, Holmkvist J, Jorgensen T, Borch-Johnsen K, Egerod KL, Lauritzen T, Sorensen TI, Bonnefond A, Meyre D, Froguel P, Schwartz TW, Pedersen O, Hansen T. MTNR1B G24E variant associates with BMI and fasting plasma glucose in the general population in studies of 22,142 Europeans. Diabetes. 2010;59:1539-48. 16. Peter I, McCaffery JM, Kelley-Hedgepeth A, Hakonarson H, Reis S, Wagenknecht LE, Kopin AS, Huggins GS, Genetics Subgroup of the Look AHEAD Study. Association of type 2 diabetes susceptibility loci with one-year weight loss in the Look AHEAD clinical trial. Obesity (Silver Spring). 2012;20:1675-82. 17. Mirzaei K, Xu M, Qi Q, de Jonge L, Bray GA, Sacks F, Qi L. Variants in glucose- and circadian rhythm-related genes affect the response of energy expenditure to weight-loss diets: The POUNDS LOST trial. Am J Clin Nutr. 2014;99:392-9. 18. Mulder H, Nagorny CL, Lyssenko V, Groop L. Melatonin receptors in pancreatic islets: Good morning to a novel type 2 diabetes gene. Diabetologia. 2009;52:1240-9.

244 Results

19. Zhang C, Bao W, Rong Y, Yang H, Bowers K, Yeung E, Kiely M. Genetic variants and the risk of gestational diabetes mellitus: A systematic review. Hum Reprod Update. 2013;19:376-90. 20. de Luis DA, Aller R, Conde R, Izaola O, Gonzalez Sagrado M, Castrodeza Sanz J. The rs9939609 gene variant in FTO modified the metabolic response of weight loss after a 3-month intervention with a hypocaloric diet. J Investig Med. 2013;61:22-6. 21. Zlatohlavek L, Vrablik M, Motykova E, Ceska R, Vasickova L, Dlouha D, Hubacek JA. FTO and MC4R gene variants determine BMI changes in children after intensive lifestyle intervention. Clin Biochem. 2013;46:313-6. 22. Goni L, Cuervo M, Milagro FI, Martinez JA. A genetic risk tool for obesity predisposition assessment and personalized nutrition implementation based on macronutrient intake. Genes Nutr. 2015;10:445. 23. Panel on Macronutrients, Panel on the Definition of Dietary Fiber, Subcommittee on Upper Reference Levels of Nutrients, Subcommittee on Interpretation and Uses of Dietary Reference Intakes, and the Standing Committee on the Scientific Evaluation of Dietary Reference Intakes. Dietary reference intakes for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein, and amino acids (macronutrients). Washington, D.C.: The National Academies Press; 2005. 24. Mataix J. Tablas de composición de alimentos. 5th ed. Granada: eug; 2009. 25. Moreiras O. Tablas de composición de alimentos. 15th ed. Madrid: Pirámide; 2012. 26. Dunbar SA. Applications of luminex xMAP technology for rapid, high-throughput multiplexed nucleic acid detection. Clin Chim Acta. 2006;363:71-82. 27. Holzapfel C, Siegrist M, Rank M, Langhof H, Grallert H, Baumert J, Irimie C, Klopp N, Wolfarth B, Illig T, Hauner H, Halle M. Association of a MTNR1B gene variant with fasting glucose and HOMA-B in children and adolescents with high BMI-SDS. Eur J Endocrinol. 2011;164:205-12. 28. Sparso T, Bonnefond A, Andersson E, Bouatia-Naji N, Holmkvist J, Wegner L, Grarup N, Gjesing AP, Banasik K, Cavalcanti-Proenca C, Marchand M, Vaxillaire M, Charpentier G, Jarvelin MR, Tichet J, Balkau B, Marre M, Levy- Marchal C, Faerch K, Borch-Johnsen K, Jorgensen T, Madsbad S, Poulsen P, Vaag A, Dina C, Hansen T, Pedersen O, Froguel P. G-allele of intronic rs10830963 in MTNR1B confers increased risk of impaired fasting glycemia and type 2 diabetes through an impaired glucose-stimulated insulin release: Studies involving 19,605 Europeans. Diabetes. 2009;58:1450-6. 29. Srinivasan V, Ohta Y, Espino J, Pariente JA, Rodriguez AB, Mohamed M, Zakaria R. Metabolic syndrome, its pathophysiology and the role of melatonin. Recent Pat Endocr Metab Immune Drug Discov. 2013;7:11-25. 30. Lyssenko V, Nagorny CL, Erdos MR, Wierup N, Jonsson A, Spegel P, Bugliani M, Saxena R, Fex M, Pulizzi N, Isomaa B, Tuomi T, Nilsson P, Kuusisto J, Tuomilehto J, Boehnke M, Altshuler D, Sundler F, Eriksson JG, Jackson AU, Laakso M, Marchetti P, Watanabe RM, Mulder H, Groop L. Common variant in MTNR1B associated with increased risk of type 2 diabetes and impaired early insulin secretion. Nat Genet. 2009;41:82-8. 31. Xia Q, Chen ZX, Wang YC, Ma YS, Zhang F, Che W, Fu D, Wang XF. Association between the melatonin receptor 1B gene polymorphism on the risk of type 2 diabetes, impaired glucose regulation: A meta-analysis. PLoS One. 2012;7:e50107. 32. Ochner CN, Barrios DM, Lee CD, Pi-Sunyer FX. Biological mechanisms that promote weight regain following weight loss in obese humans. Physiol Behav. 2013;120:106-13. 33. van Nas A, Guhathakurta D, Wang SS, Yehya N, Horvath S, Zhang B, Ingram-Drake L, Chaudhuri G, Schadt EE, Drake TA, Arnold AP, Lusis AJ. Elucidating the role of gonadal hormones in sexually dimorphic gene coexpression networks. Endocrinology. 2009;150:1235-49.

245 Chapter 7

34. Dempfle A, Hinney A, Heinzel-Gutenbrunner M, Raab M, Geller F, Gudermann T, Schafer H, Hebebrand J. Large quantitative effect of melanocortin-4 receptor gene mutations on body mass index. J Med Genet. 2004;41:795-800. 35. Lindgren CM, Heid IM, Randall JC, Lamina C, Steinthorsdottir V, Qi L, et al. Genome-wide association scan meta- analysis identifies three loci influencing adiposity and fat distribution. PLoS Genet. 2009;5:e1000508. 36. Heid IM, Jackson AU, Randall JC, Winkler TW, Qi L, Steinthorsdottir V, et al. Meta-analysis identifies 13 new loci associated with waist-hip ratio and reveals sexual dimorphism in the genetic basis of fat distribution. Nat Genet. 2010;42:949-60. 37. Randall JC, Winkler TW, Kutalik Z, Berndt SI, Jackson AU, Monda KL, et al. Sex-stratified genome-wide association studies including 270,000 individuals show sexual dimorphism in genetic loci for anthropometric traits. PLoS Genet. 2013;9:e1003500. 38. Navas-Carretero S, Holst C, Saris W, van Baak M, Jebb S, Kafatos A, Papadaki A, Pfeiffer A, Handjieva-Darlenska T, Hlavaty P, Stender S, Larsen T, Astrup A, Martinez J. Gender and protein intake impact on the success of weight maintenance and cardiovascular risk benefits, independently of the mode of food provision: The DiOGenes randomised trial. J Am Coll Nutr. In press. 39. Loos RJ, Yeo GS. The bigger picture of FTO: The first GWAS-identified obesity gene. Nat Rev Endocrinol. 2014;10:51-61. 40. Villalobos-Comparan M, Teresa Flores-Dorantes M, Teresa Villarreal-Molina M, Rodriguez-Cruz M, Garcia-Ulloa AC, Robles L, Huertas-Vazquez A, Saucedo-Villarreal N, Lopez-Alarcon M, Sanchez-Munoz F, Dominguez-Lopez A, Gutierrez-Aguilar R, Menjivar M, Coral-Vazquez R, Hernandez-Stengele G, Vital-Reyes VS, Acuna-Alonzo V, Romero- Hidalgo S, Ruiz-Gomez DG, Riano-Barros D, Herrera MF, Gomez-Perez FJ, Froguel P, Garcia-Garcia E, Teresa Tusie- Luna M, Aguilar-Salinas CA, Canizales-Quinteros S. The FTO gene is associated with adulthood obesity in the Mexican population. Obesity (Silver Spring). 2008;16:2296-301. 41. Lappalainen T, Kolehmainen M, Schwab U, Pulkkinen L, de Mello VD, Vaittinen M, Laaksonen DE, Poutanen K, Uusitupa M, Gylling H. Gene expression of FTO in human subcutaneous adipose tissue, peripheral blood mononuclear cells and adipocyte cell line. J Nutrigenet Nutrigenomics. 2010;3:37-45. 42. Smemo S, Tena JJ, Kim KH, Gamazon ER, Sakabe NJ, Gomez-Marin C, Aneas I, Credidio FL, Sobreira DR, Wasserman NF, Lee JH, Puviindran V, Tam D, Shen M, Son JE, Vakili NA, Sung HK, Naranjo S, Acemel RD, Manzanares M, Nagy A, Cox NJ, Hui CC, Gomez-Skarmeta JL, Nobrega MA. Obesity-associated variants within FTO form long- range functional connections with IRX3. Nature. 2014;507:371-5. 43. Loos RJ. The genetic epidemiology of melanocortin 4 receptor variants. Eur J Pharmacol. 2011;660:156-64. 44. Hinney A, Volckmar AL, Knoll N. Melanocortin-4 receptor in energy homeostasis and obesity pathogenesis. Prog Mol Biol Transl Sci. 2013;114:147-91. 45. El-Sayed Moustafa JS, Froguel P. From obesity genetics to the future of personalized obesity therapy. Nat Rev Endocrinol. 2013;9:402-13. 46. Zhang X, Qi Q, Zhang C, Smith SR, Hu FB, Sacks FM, Bray GA, Qi L. FTO genotype and 2-year change in body composition and fat distribution in response to weight-loss diets: The POUNDS LOST trial. Diabetes. 2012;61:3005- 11. 47. Verhoef SP, Camps SG, Bouwman FG, Mariman EC, Westerterp KR. Genetic predisposition, dietary restraint and disinhibition in relation to short and long-term weight loss. Physiol Behav. 2014;128:247-51. 48. Qi L. Gene-diet interaction and weight loss. Curr Opin Lipidol. 2014;25:27-34. 49. Martinez JA, Navas-Carretero S, Saris WH, Astrup A. Personalized weight loss strategies-the role of macronutrient distribution. Nat Rev Endocrinol. 2014.

246 Results

50. Oishi K, Uchida D, Itoh N. Low-carbohydrate, high-protein diet affects rhythmic expression of gluconeogenic regulatory and circadian clock genes in mouse peripheral tissues. Chronobiol Int. 2012;29:799-809. 51. Teixeira PJ, Going SB, Houtkooper LB, Cussler EC, Metcalfe LL, Blew RM, Sardinha LB, Lohman TG. Pretreatment predictors of attrition and successful weight management in women. Int J Obes Relat Metab Disord. 2004;28:1124- 33. 52. Westerterp-Plantenga MS, Nieuwenhuizen A, Tome D, Soenen S, Westerterp KR. Dietary protein, weight loss, and weight maintenance. Annu Rev Nutr. 2009;29:21-41. 53. Ortner Hadziabdic M, Mucalo I, Hrabac P, Matic T, Rahelic D, Bozikov V. Factors predictive of drop-out and weight loss success in weight management of obese patients. J Hum Nutr Diet. 2014;13.doi: 10.1111/jhn.12270.

247 Chapter 7

Table 1. Baseline characteristics according to MTNR1B rs10830963 genotype

Total CC CG/GG p-value (mean ±SE) (mean ±SE) (mean ±SE) n (%) 167 88 (52.7) 79 (47.3) - Women (%) 126 (75.4) 65 (73.9) 61 (77.2) 0.252a Age (years) b 51.9 ± 1.0 52.7 ± 1.4 51.1 ± 1.5 0.440 Anthropometric measurements Height (cm) b 163.2 ± 0.7 162.7 ± 0.6 163.8 ± 0.7 0.261 Weight (kg) b 84.2 ± 1.2 84.7 ± 1.4 83.7 ± 1.5 0.625 BMI (kg/m2) b 31.6 ± 0.4 31.9 ± 0.5 31.1 ± 0.5 0.270 BFM (%) b 37.6 ± 0.7 37.8 ± 0.7 37.4 ± 0.7 0.637 Waist circumference (cm) b 101.4 ± 1.1 102.3 ± 1.2 100.4 ± 1.3 0.286 Hip circumference (cm) b 113.0 ± 0.8 113.8 ± 1.0 112.1 ± 1.0 0.255 Waist/ hip (ratio) b 0.90 ± 0.01 0.90 ± 0.01 0.89 ± 0.01 0.783 Waist/ height (ratio) b 0.62 ± 0.01 0.63 ± 0.01 0.61 ± 0.01 0.156 Physical activity Physical activity level c 1.23 ± 0.00 1.23 ± 0.00 1.23 ± 0.00 0.570 Baseline dietary intake Energy (kcal/day) d 2,230 ± 33 2,212 ± 41 2,249 ± 43 0.536 Carbohydrate (%E) b 36.1 ± 0.8 36.1 ± 1.1 36.0 ± 1.1 0.961 Fat (%E) b 40.6 ± 0.6 40.2 ± 0.8 41.0 ± 0.8 0.458 Protein (%E) b 17.0 ± 0.2 17.2 ± 0.3 16.7 ± 0.3 0.338 Animal protein (g/day) b 61.6 ± 1.6 62.5 ± 2.0 60.5 ± 2.2 0.516 Vegetable protein (g/day) b 32.5 ± 0.7 32.5 ± 0.6 32.6 ± 0.7 0.960 Genetic variants rs9939609 (FTO) TT 51 (30.5) 28 (31.8) 23 (29.1) 0.918a TA 82 (49.1) 42 (47.7) 40 (50.6) AA 34 (20.4) 18 (20.5) 16 (20.3) rs17782313 (MC4R) TT 99 (59.3) 53 (60.2) 46 (58.2) 0.963a TC 62 (37.1) 32 (36.4) 30 (38.0) CC 6 (3.6) 3 (3.4) 3 (3.8) a χ2 test p-value bAdjusted for gender, age, energy intake and physical activity cAdjusted for gender, age and energy intake dAdjusted for gender, age, and physical activity

248 Results

Table 2. Linear regression coefficients for the association between MTNR1B and anthropometric variables after weight loss intervention

Δ weight Δ BMI B (95% CI) p B (95% CI) p All CC 0 (ref.) 0 (ref.) CG/GG -0.11 (-0.64, 0.42) 0.686 -0.08 (-0.27, 0.11) 0.391 Males CC 0 (ref.) 0 (ref.) CG/GG 0.73 (-1.05, 2.51) 0.409 0.24 (-0.30, 0.79) 0.368 Females CC 0 (ref.) 0 (ref.) CG/GG -0.55 (-1.07, -0.03) 0.040 -0.23 (-0.45, -0.03) 0.027 Adjusted for gender, age, energy restriction, time between visits 1 and 2, FTO and MC4R variants and baseline value of the appropriate anthropometric variable

249 Chapter 7

Figure 1. Influence of obesity loci on the association of MTNR1B and BMI loss in the total population Adjusted for age, energy restriction, time between visits, and BMI at baseline

250 Results

Figure 2. Interaction between MTRN1B and total protein intake at baseline (A) and animal protein intake at baseline (B) on body weight loss, among women Adjusted for age, energy restriction, time between visits, FTO, MC4R and body weight at baseline.

251

CHAPTER 8

Circadian rhythms related MTNR1B genetic variant modulates the effect of weight-loss diets on changes in adiposity and body composition: The POUNDS Lost trial

Goni L.1,2, Sun D.3, Heianza Y.3, Wang T.3, Huang T.4, Martínez J.A.1,2,5,6, Shang X.7, Bray G.A.8, Smith S.R.8, Sacks F.M.9, Qi L.3,9,10,11

1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Navarra, Spain 2 Centre for Nutrition Research, University of Navarra, Navarra, Spain 3 Department of Epidemiology, Tulane University, LA, USA 4 Epidemiology Domain, Saw Swee Hoch School of Public Health and Department of Medicine, National University of Singapore, Singapore, Singapore 5 Navarra Institute for Health Research (IdiSNA), Navarra, Spain 6 Biomedical Research Centre Network in Physiopathology of Obesity and Nutrition (CIBERobn), Institute of Health Carlos III, Madrid, Spain 7 Children’s Hospital New Orleans, LA, USA 8 Pennington Biomedical Research Center, Louisiana State University, LA, USA 9 Department of Nutrition, Harvard T.H. Chan School of Public Health, MA, USA 10 Department of Epidemiology, Harvard T.H. Chan School of Public Health, MA, USA 11 Channing Laboratory, Brigham and Women’s Hospital and Harvard Medical School, MA, USA

Eur J Nutr Under review (EJON-D-17-00512) Impact factor (2016): 4.370 14/81 Nutrition & Dietetics, Q1

Results

ABSTRACT

Purpose. Melatonin receptor 1B (MTNR1B) gene is a circadian rhythms associated common variant which is related to increased signaling of melatonin, a hormone previously related to body fatness mainly through effects on energy metabolism. We examined whether the MTNR1B variant affects changes of body fatness and composition in response to a dietary weight loss intervention.

Methods. The MTNR1B rs10830963 variant was genotyped for 722 overweight and obese individuals, who were randomly assigned to one of four diets varying in macronutrient composition. Anthropometric and body composition measurements (DXA scan) were collected at baseline and at 6 and 24 months of follow-up.

Results. Statistically significant interactions were observed between the MTNR1B genotype and low-/high-fat diet on changes in weight, body mass index (BMI), waist circumference (WC) and total body fat (p interaction=0.01, 0.02, 0.002 and 0.04, respectively), at 6 months of dietary intervention. In the low-fat diet group, increasing number of the sleep disruption related G allele was significantly associated with a decrease in weight (p=0.004), BMI (p=0.005) and WC (p=0.001). Whereas, in the high-fat diet group, carrying the G allele was associated with an increase in body fat (p=0.03). At 2 years, the associations remained statistically significant for changes in body weight (p=0.02), BMI (p=0.02) and WC (p=0.048) in the low-fat diet group, although the gene-diet interaction became less significant.

Conclusions. The results suggest that carriers of the G allele of the MTNR1B rs10830963 may have a greater improvement in body adiposity and fat distribution when eating a low-fat diet.

Keywords

Melatonin receptor 1B; gene-diet interaction; high-fat diet; weight-loss intervention; adiposity

255 Chapter 8

1. INTRODUCTION

It has long been recognized that circadian system is implicated in the regulation of energy balance, and subsequently affecting body fatness [1]. In humans beings, disruption of circadian rhythms by means of shift work, social jet lag, sleep deprivation, timed feeding and consumption of a high-fat diet among others, has been related to obesity and metabolic disturbances such as type 2 diabetes and cardiovascular disease [2–4]. One of the most important chronobiotics is melatonin, a hormone secreted mainly by the pineal gland and a key mediator used by the central master clock to synchronize the circadian system [5].

Recent genome-wide association studies (GWAS) have identified common variants in the Melatonin receptor 1B (MTNR1B) gene, which encodes one of the two high-affinity receptors for melatonin [6, 7], associated with fasting plasma glucose and the risk of type 2 diabetes [8– 10]. Among them, the MTNR1B rs10830963 risk allele has been related to melatonin signaling [11]. The sleep disruption G allele has also been associated with adiposity measures such as BMI and waist circumference (WC) as well as with body weight loss [12–15]. Notably, evidence has indicated that melatonin also plays a key role in the regulation of adipocyte biology (lipolysis, lipogenesis), the activation of brown adipose tissue and participation in the browning process of white adipose tissue, and the maintenance of an adequate energy balance acting on the regulation of energy expenditure and energy intake [5, 16, 17]. Therefore, we hypothesized that the MTNR1B genotype might affect changes in body fatness and composition in response to dietary interventions.

In the current study, we investigated the effect of the MTNR1B rs10830963 genetic variant on changes in body fatness and body composition in response to weight-loss diets varying in macronutrient contents in the Preventing Overweight Using Novel Strategies (POUNDS Lost) trial.

2. METHODS

2.1. Study population

The POUNDS Lost trial is a 2-year randomized clinical trial (clinical trial reg. no. NCT00072995) conducted from October 2004 through December 2007 at two sites: Harvard School of Public Health and Brigham and Women’s Hospital in Boston, MA, and the Pennington Biomedical Research Center of Louisiana State University System, Baton Rouge, LA. Details of the study design and methods have been described elsewhere [18]. This 2-year study included 811 overweight and obese (BMI 25–40 kg/m2) individuals who were randomly assigned to one of

256 Results four diets in which calories were restricted by 750 kcal/day. The target percentages of energy derived from fat, protein and carbohydrate in the four diets were: 20, 15 and 60%; 20, 25 and 55%; 40, 15 and 45%; and 40, 25 and 35%, respectively. Thus, the 4 diets constituted a 2-by-2 factorial design: two diets were low fat (20%), two diets were high fat (40%), two diets were average protein (15%) and two diets were high protein (25%). Major exclusion criteria in this trial were the presence of diabetes or unstable cardiovascular disease, the use of medications that affect body weight, and insufficient motivation [18]. The study was approved by the human subjects committee at each institution and by a data and safety monitoring board appointed by the National Heart, Lung and Blood Institute. All participants provided written informed consent.

In the present study, 722 subjects with baseline genotyped data of the MTNR1B rs10830963 variant, were included 19. Among them, 79.3% were white, 15.7% were African American and 5.0% were Hispanic or other ethnic groups by self-report. For the analyses, body weight and WC data were available for 722 individuals at baseline, 648 and 645 individuals at 6 months, and 587 and 549 individuals at 2 years, respectively. A dual-energy X-ray absorptiometry (DXA) scan was carried out in a random sample of 50% of the total study population, including 382 participants at baseline, 304 participants at 6 months and 222 participants at 2 years.

2.2. Measurements

Body weight and WC were measured in the morning before breakfast at baseline, 6 months and 2 years. Body weight was measured by calibrated hospital scales and WC was measured using a non-stretchable tape measure, 4 cm above the iliac crest. Height was measured at the baseline examination. BMI was calculated dividing weight (kg) by the square of height (m2). Body composition was analyzed by a DXA scan using a Hologic QDR 4500A (Hologic, Inc. Waltham, MA) after an overnight fast [20]. Total fat mass (kg), total lean mass (kg), the percentage of whole body fat mass and percentage of trunk fat were obtained at baseline, 6 months and 2 years of the intervention. To evaluate the adherence to the dietary intervention program, dietary intake was assessed in a random sample of 50% of the participants by a review of the 5 days diet record at baseline and by 24-h recall during a telephone interview on 3 nonconsecutive days at 6 months and 2 years of follow-up.

257 Chapter 8

2.3. Genotyping

DNA was isolated from the buffy coat fraction of centrifuged blood using the QIAmp Blood Kit (Qiagen). The previously reported single nucleotide polymorphism (SNP) MTNR1B rs10830963 was successfully genotyped in 722 of 811 total participants using the OpenArray SNP Genotyping System (BioTrove) [19]. Replicated quality control samples (10%) were included and genotyped with >99% concordance. The genotype distribution was consistent with Hardy- Weinberg equilibrium (HWE) in all study participants or in the major ethnic group (whites) (p > 0.05).

2.4. Statistical analyses

The primary end points of this study were changes in body fatness (weight, BMI, WC) and composition measures (body fat, lean body mass, total fat mass percentage and trunk fat percentage) over the course of the intervention. Chi-squared test for categorical variables, and general linear models for continuous variables were performed for comparison of baseline characteristics across genotypes. To compare changes in body fatness and body composition measurements by genotypes according to low- or high-fat group at 6 months and at 2 years of the diet intervention, general linear models were used adjusted for covariates (Model 1 adjusted for age, gender, ethnicity and the respective baseline variable; Model 2 adjusted for model 1 plus BMI at baseline). Gene-diet interactions were evaluated including the interaction term in the models (e.g., MTNR1B genotype x high-/low-fat diet group). Sensitivity analyses were performed among white individuals to evaluate the influence of potential population stratification. Additive genetic models were used in the analysis. Statistical analyses were performed using STATA/SE version 12.0 (StataCorp, College Station, TX, USA). A p value < 0.05 was considered statistically significant.

3. RESULTS

Table 1 shows baseline characteristics of the participants according to the MTNR1B rs10830963 genotype. The minor allele frequency (G allele) was 0.27 in the study population. The genotype frequencies were similar between males and females and among diet groups. Nonetheless, the distribution of the SNP was different by ethnicity (p<0.001). Dietary intake and body fatness and compositional measurements were not related to the MTNR1B genetic variant at baseline examination. There were no associations of genotype with changes in body fatness and compositional measurements at 6 months and 2 years of diet intervention, after

258 Results adjustment for age, sex ethnicity, BMI at baseline (if appropriate), high-/low-fat diet group and baseline value for the according outcomes, were observed (data not shown) [19].

Although the targets of macronutrient intakes were not fully achieved, the reported dietary intake and changes in adherence biomarkers confirmed that participants modified their intakes of macronutrients in the direction of the intervention (Supplementary Table 1). There were no significant differences in mean values of macronutrient intakes and urinary protein at 6 months and 2 years across the MTNR1B rs10830963 genotype and group diet, except for fat intake at 24 months among participants in the high-fat group (p=0.04). Moreover, the G allele was significantly associated with a greater increase in RQ in the low-fat diet group at 24 months, as has been previously reported [19].

After adjustment for age, sex, ethnicity, BMI at baseline (if appropriate) and the respective baseline variable, significant interactions between the MTNR1B rs10830963 genotype and high-/low-fat diet on changes in weight, BMI and WC were observed (Table 2 and Figure 1). An increasing number of the G allele of the MTNR1B genetic variant was significantly associated with a decrease in body weight, BMI and WC at 6 months in response to the low-fat diet (p interaction=0.01, 0.02 and 0.002, respectively). When we applied the analyses among whites similar trends with significant interactions at 6 months on changes in body weight, BMI and WC were found (all p interaction<0.05).

The interaction between the MTNR1B rs10830963 genotypes and dietary fat was also analyzed for the measurements of body composition including total fat, total lean body mass, percentage total fat mass and percentage trunk fat at 6 months (Table 2 and Figure 1). Consistent with the observations of changes in body fatness measurements, a statistically significant interaction between the MTNR1B genetic variant and low-/high-fat diet was found on changes in total fat, adjusting for age, gender, ethnicity, BMI at baseline and the respective baseline variable (p interaction=0.04). In response to the high-fat diet, an increasing number of the G allele was significantly associated with an increase of total fat (p=0.03). Similar trends were also found for percentage total fat mass and percentage trunk fat although the gene- dietary fat interaction was no longer statistically significant (p interaction=0.07 and 0.10, respectively) after adjustment for covariates (model 2). Similar results were observed in the white population (data not were shown), although the interaction for body fat was not statistically significant (p interaction=0.07, model 2).

At 2 years participants, on average, regained body weight [18]. The associations with MTNR1B rs10830963 genotypes remained statistically significant for weight (p=0.02), BMI (p=0.02) and

259 Chapter 8

WC (p=0.048) in the low-fat diet group, although the gene-diet interaction became non- significant (Table 2 and Figure 2).

4. DISCUSSION

In this 2-year randomized dietary weight-loss intervention trial, we found significant interactions between the circadian rhythm genotype MTNR1B and dietary fat intake on changes in fatness, fat distribution and body composition measurements, especially at 6 months. In response to the low-fat diet, increasing number of the G allele was associated with a greater reduction in body weight, BMI, WC and total body fat.

The circadian system has long been implicated in the regulation of body fatness due to its role in the control of energy balance [21]. Mice with mutations in the circadian rhythm gene Clock fed with a high-fat diet developed obesity at a young age as wells as a variety of metabolic and endocrine abnormalities consistent with the metabolic syndrome [22]. In addition, the circadian Clock mutant mice exhibited decreased expression of transcripts encoding selected hypothalamic peptides involved in energy balance. Melatonin is used by the central master clock as an internal synchronizer coordinating central and peripheral tissues [5]. Our findings are consistent with the biological role of melatonin in energy metabolism and energy balance [5]. On one hand, melatonin plays roles not only in the regulation of metabolic processes but also in the maintenance of their circadian organization [23]. On the other hand, the effect of melatonin on energy balance has been consistently observed [5]. Wolden-Hanson et al. demonstrated that melatonin supplementation therapy decreased body weight and intraabdominal fat, and increased the nocturnal locomotor activity and core body temperature [24]. While the precise mechanisms underlying our results remain largely unknown, several lines of evidence have implicated the MTNR1B rs10830963 genotype in regulating melatonin signaling [11]. Tuomi et al. ascertained that subjects carrying 1 and 2 MTNR1B rs10830963 G alleles showed a 2- and 4-fold increase in MTNR1B mRNA expression in human pancreatic islets, respectively, compared with the non-carriers [11]. Moreover, the authors demonstrated that the administration of melatonin to nondiabetic individuals inhibits insulin secretion in all subjects, and the effect was stronger among GG than those did not carry this allele. Another recent study reported that the common genetic variant was associated with the timing of the melatonin rhythm [25]. MTNR1B rs10830963 G allele carriers showed a later melatonin offset and a longer duration of elevated melatonin levels. The authors suggested that the disruption of melatonin rhythm among carriers of the risk allele may result in an increase of food intake

260 Results to coincide with elevated melatonin levels in the morning leading to decreased glucose tolerance.

In line with our previous results, we found that the relations of the MTNR1B rs10830963 genetic variant with changes in fatness and body composition measurements were significantly modified by dietary fat intake [19]. Notably, a high-fat diet has been found to alter the circadian molecular clock, both centrally and peripherally leading to a state of chronodisruption [1, 26]. Interestingly, several previous studies reported that the expression levels of clock genes and also the rhythmic mRNA expression were influenced by a high-fat diet consumption, in different animal tissues [27, 28]. Moreover, it has been demonstrated that a significant alteration of circadian rhythmicity of different hormones is related to obesity such as pineal melatonin, leptin, ghrelin and adiponectin among others, in rats fed a high-fat diet [29–31]. Taken together, these data lent support to potential interplays between the MTNR1B genotype and dietary fat.

Our findings of body composition analysis suggested that the MTNR1B genotype might affect total body fat composition, instead of specific fat compositions (trunk fat). According to our results, a previous study showed that MTNR1B was expressed in rodent inguinal and epididymal adipocytes [32]. In this sense, it should be highlighted, that the effect of melatonin on adipocyte biology is not only mediated by means of the receptors in adipocytes but also through the action of the sympathetic nervous system [16]. In humans, Staiger et al. found an association between the MTNR1B rs4753426 genetic variant and total body fat [33]. More studies are needed to further verify our findings.

In the current study, the gene-diet interactions on changes in body fatness and composition measurements were attenuated at 2 years of follow-up. On the one hand, this might be due in part to decreasing adherence to the diet that occurred between 6 months and 2 years in the POUNDS Lost trial, similar to other weight-loss interventions [18, 34–36]. On the other hand, the statistical power might also be diminished because of more dropouts at 2 years of diet intervention than at 6 months. Even though, similar relations of the MTNR1B rs10830963 genotype with a decrease in body weight, BMI and WC in the low-fat diet group were observed at 2 years, indicating the genetic effects were stable and for up to 2 years.

As far as we know, this is the first study to date to report significant interactions between the MTNR1B rs10830963 genetic variant and dietary fat intake on changes in fatness and body compositional measurements in a large and long-term randomized weight-loss dietary intervention trial. Nonetheless, we acknowledge several limitations. First, we did not measure

261 Chapter 8 clock parameters such as melatonin and MTNR1B expression which limited our ability to explore plausible underlying mechanisms. Second, the power to detect long-term genotypic effect in response to a high-/low-fat diet was reduced due to the decline of adherence after 6 months of diet intervention, as has been reported by other authors [18, 34–36]. Third, because low-fat intake is usually characterized by high-carbohydrate intake, it is difficult to determine which macronutrient would best explain our results. Finally, around 80% of the study participants were white, and the genotype distribution differed across ethnicity, so further studies are required to generalize our findings to other ethnic groups.

In conclusion, the results of the present study showed that the level of fat intake might modify the effect of the circadian rhythms related MTNR1B rs10830963 genetic variant on changes in body fatness and composition. Subjects with the GG genotype might have a better response to a weight-loss dietary intervention by choosing a low-fat diet.

Ethical standards

All procedures of the present study were in accordance with the ethical standards laid down the 1964 Helsinki declaration and its later amendments.

The study was approved by the human subjects committee at each institution and by a data and safety monitoring board appointed by the National Heart, Lung and Blood Institute.

All participants provided written informed consent.

Conflict of interest

The authors declare that they have no conflict of interest.

262 Results

REFERENCES

1. Laermans J, Depoortere I (2016) Chronobesity: Role of the circadian system in the obesity epidemic. Obes Rev 17:108–125. doi: 10.1111/obr.12351 2. Garaulet M, Ordovás JM, Madrid JA (2010) The chronobiology, etiology and pathophysiology of obesity. Int J Obes (Lond) 34:1667–1683. doi: 10.1038/ijo.2010.118 3. Qian J, Scheer FAJL (2016) Circadian system and glucose metabolism: Implications for physiology and disease. Trends Endocrinol Metab 27:282–293. doi: 10.1016/j.tem.2016.03.005 4. Scott EM (2015) Circadian clocks, obesity and cardiometabolic function. Diabetes, Obes Metab 17:84–89. doi: 10.1111/dom.12518 5. Cipolla-Neto J, Amaral FG, Afeche SC, et al (2014) Melatonin, energy metabolism, and obesity: A review. J Pineal Res 56:371–381. doi: 10.1111/jpi.12137 6. Dubocovich ML, Delagrange P, Krause DN, et al (2010) Nomenclature, classification, and pharmacology of G Protein-Coupled Melatonin Receptors. Pharmocological Rev 62:343–380. doi: 10.1124/pr.110.002832.343 7. Gamble KL, Berry R, Frank SJ, Young ME (2014) Circadian clock control of endocrine factors. Nat Rev Endocrinol 10:466–475. doi: 10.1038/nrendo.2014.78 8. Prokopenko I, Langenberg C, Florez JC, et al (2009) Variants in MTNR1B influence fasting glucose levels. Nat Genet 41:77–81. doi: 10.1038/ng.290 9. Manning AK, Hivert M-F, Scott RA, et al (2012) A genome-wide approach accounting for body mass index identifies genetic variants influencing fasting glycemic traits and insulin resistance. Nat Genet 44:659–69. doi: 10.1038/ng.2274 10. Gaulton KJ, Ferreira T, Lee Y, et al (2015) Genetic fine mapping and genomic annotation defines causal mechanisms at type 2 diabetes susceptibility loci. Nat Genet 47:1415–25. doi: 10.1038/ng.3437 11. Tuomi T, Nagorny CLF, Singh P, et al (2016) Increased melatonin signaling is a risk factor for type 2 diabetes. Cell Metab 23:1067–1077. doi: 10.1016/j.cmet.2016.04.009 12. Stancáková A, Kuulasmaa T, Paananen J, et al (2009) Association of 18 confirmed susceptibility loci for type 2 diabetes with indices of insulin release, proinsulin conversion, and insulin sensitivity in 5,327 nondiabetic Finnish men. Diabetes 58:2129–2136. 13. Kong X, Zhang X, Xing X, et al (2015) The association of type 2 diabetes loci identified in genome-wide association studies with metabolic syndrome and its components in a Chinese population with type 2 diabetes. PLoS One 10:1–21. doi: 10.1371/journal.pone.0143607 14. Grotenfelt NE, Wasenius NS, Rönö K, et al (2016) Interaction between rs10830963 polymorphism in MTNR1B and lifestyle intervention on occurrence of gestational diabetes. Diabetologia 59:1655–1658. doi: 10.1007/s00125- 016-3989-1 15. Goni L, Cuervo M, Milagro FI, Martínez JA (2014) Gene-gene interplay and gene-diet interactions involving the MTNR1B rs10830963 variant with body weight loss. J Nutrigenet Nutrigenomics 7:232–242. 16. Szewczyk-Golec K, Woźniak A, Reiter RJ (2015) Inter-relationships of the chronobiotic, melatonin, with leptin and adiponectin: Implications for obesity. J Pineal Res 59:277–291. doi: 10.1111/jpi.12257 17. Barrenetxe J, Delagrange P, Martínez JA (2004) Physiological and metabolic functions of melatonin. J Physiol Biochem 60:61–72. 18. Sacks FM, Bray GA, Carey VJ, et al (2009) Comparison of weight-loss diets with different compositions of fat, protein, and carbohydrates. N Engl J Med 360:859–873. doi: 10.1056/NEJMoa1208410

263 Chapter 8

19. Mirzaei K, Xu M, Qi Q, et al (2014) Variants in glucose- and circadian rhythm-related genes affect the response of energy expenditure to weight-loss diets: the POUNDS LOST Trial. Am J Clin Nutr 99:392–399. 20. De Souza RJ, Bray GA, Carey VJ, et al (2012) Effects of 4 weight-loss diets differing in fat, protein, and carbohydrate on fat mass, lean mass, visceral adipose tissue, and hepatic fat: Results from the POUNDS LOST trial. Am J Clin Nutr 95:614–625. doi: 10.3945/ajcn.111.026328 21. Summa KC, Turek FW (2014) Chronobiology and obesity: Interactions between circadian rhythms and energy regulation. Adv Nutr 5:312S–9S. doi: 10.3945/an.113.005132 22. Turek FW, Joshu C, Kohsaka A, et al (2005) Obesity and metabolic syndrome in circadian Clock mutant mice. Science (80- ) 308:1043–1045. doi: 10.1126/science.1108750 23. Kumar Jha P, Challet E, Kalsbeek A (2015) Circadian rhythms in glucose and lipid metabolism in nocturnal and diurnal mammals. Mol Cell Endocrinol 418:74–88. doi: 10.1016/j.mce.2015.01.024 24. Wolden-Hanson T, Mitton DR, McCants RL, et al (2000) Daily melatonin administration to middle-aged male rats suppresses body weight, intraabdominal adiposity, and plasma leptin and insulin independent of food intake and total body fat. Endocrinology 141:487–497. doi: 10.1210/en.141.2.487 25. Lane JM, Chang AM, Bjonnes AC, et al (2016) Impact of common diabetes risk variant in MTNR1B on sleep, circadian, and melatonin physiology. Diabetes 65:1741–1751. doi: 10.2337/db15-0999 26. Oosterman JE, Kalsbeek A, la Fleur SE, Belsham DD (2015) Impact of nutrients on circadian rhythmicity. Am J Physiol - Regul Integr Comp Physiol 308:R337–R350. doi: 10.1152/ajpregu.00322.2014 27. Yanagihara H, Ando H, Hayashi Y, et al (2006) High-fat feeding exerts minimal effects on rhythmic mRNA expression of clock genes in mouse peripheral tissues. Chronobiol Int 23:905–914. doi: 10.1080/07420520600827103 28. Hsieh M-C, Yang S-C, Tseng H-L, et al (2010) Abnormal expressions of circadian-clock and circadian clock- controlled genes in the livers and kidneys of long-term, high-fat-diet-treated mice. Int J Obes (Lond) 34:227–239. doi: 10.1038/ijo.2009.228 29. Cano P, Jiménez-Ortega V, Larrad Á, et al (2008) Effect of a high-fat diet on 24-h pattern of circulating levels of prolactin, luteinizing hormone, testosterone, corticosterone, thyroid-stimulating hormone and glucose, and pineal melatonin content, in rats. Endocrine 33:118–125. doi: 10.1007/s12020-008-9066-x 30. Cano P, Cardinali DP, Rios-Lugo MJ, et al (2009) Effect of a high-fat diet on 24-hour pattern of circulating adipocytokines in rats. Obes (Silver Spring) 17:1866–1871. doi: oby2009200 [pii]\r10.1038/oby.2009.200 31. Cardinali DP, Cano P, Jiménez-Ortega V, Esquifino AI (2011) Melatonin and the metabolic syndrome: Physiopathologic and therapeutical implications. Neuroendocrinology 93:133–142. doi: 10.1159/000324699 32. Zalatan F, Krause JA, Blask DE (2001) Inhibition of isoproterenol-induced lipolysis in rat inguinal adipocytes in vitro by physiological melatonin via a receptor-mediated mechanism. Endocrinology 142:3783–3790. 33. Staiger H, Machicao F, Schäfer SA, et al (2008) Polymorphisms within the novel type 2 diabetes risk locus MTNR1B determine β-cell function. PLoS One 3:e3962. doi: 10.1371/journal.pone.0003962 34. Foster GD, Wyatt HR, Hill JO, et al (2003) A randomized trial of a low-carbohydrate diet for obesity. N Engl J Med 348:2082–2090. doi: 10.1056/NEJMoa022207 35. Dansinger ML, Gleason JA, Griffith JL, et al (2005) Comparison of the Atkins, Ornish, Weight Watchers, and Zone Diets for weight loss and heart disease risk reduction: A randomized trial. J Am Med Assoc 293:43–53. doi: 10.1097/00008483-200505000-00012 36. Shai I, Schwarzfuchs D, Henkin Y, et al (2008) Weight loss with a low-carbohydrate, Mediterranean, or low-fat diet. N Engl J Med 359:229–241. doi: 10.1056/NEJMoa1208410

264 Results

Table 1. Baseline characteristics of study participants according to the MTNR1B rs10830963 genetic variant

CC CG GG p value n=393 n=273 n=56 Age (years) 50.8 (9.2) 51.0 (9.1) 53.0 (10.3) 0.24 Sex 0.72 Male 149 (37.9) 112 (41.0) 22 (39.3) Female 244 (62.1) 161 (59.0) 34 (60.7) Race or ethnic group <0.001 White 279 (71.0) 247 (90.5) 49 (87.5) Black 96 (24.4) 12 (4.4) 3 (5.4) Hispanic or other 18 (4.6) 14 (5.1) 4 (7.1) Diet group 0.69 Low fat 198 (50.4) 134 (49.1) 31 (55.4) High fat 195 (49.6) 139 (50.2) 25 (44.6) Dietary intake per day a Energy (kcal) 1945 (555) 2016 (572) 1884 (448) 0.36 Protein (%) 18.0 (3.4) 18.3 (3.0) 18.2 (3.2) 0.77 Fat (%) 36.9 (5.9) 37.0 (6.1) 37.6 (6.3) 0.85 Carbohydrate (%) 45.1 (7.8) 43.8 (7.6) 44.9 (7.3) 0.31 Body weight (kg) 93.7 (15.0) 92.8 (16.0) 92.0 (17.5) 0.66 BMI (kg/m2) 32.9 (3.8) 32.3 (3.9) 32.5 (3.7) 0.11 WC (cm) 103.9 (13.0) 103.2 (13.2) 103.9 (13.4) 0.78 Body composition b Total body fat (kg) 35.3 (8.0) 34.5 (8.1) 34.6 (5.6) 0.62 Total body lean (kg) 60.8 (12.8) 60.4 (13.6) 56.5 (13.1) 0.26 Total body fat mass (%) 37.0 (7.1) 36.6 (6.9) 38.4 (6.5) 0.44 Trunk fat (%) 38.2 (6.2) 37.5 (6.1) 38.6 (5.8) 0.54 BMI, Body mass index; WC, Waist circumference Data are expressed as mean (SD) or n (%) a Data were available for 370 individuals (CC n=208, CG n=133, GG n=29) b Data were available for 382 individuals (CC n=204, CG n=149, GG n=29)

265 Chapter 8

Table 2. Effect of the MTNR1B rs10830963 genetic variant on changes in body weight, BMI, WC and body composition in response to low-/high-fat diet at 6 months and 2 years of diet intervention

6 months 2 years Low fat High fat Low fat High fat p interaction p interaction b p value b p value b p value b p value Model 1 D Weight (kg) -1.38 (0.48) 0.004 0.55 (0.51) 0.29 0.01 -1.60 (0.66) 0.02 -0.05 (0.68) 0.90 0.19 D BMI (kg/m2) -0.48 (0.17) 0.005 0.15 (0.18) 0.39 0.02 -0.53 (0.23) 0.02 -0.07 (0.23) 0.76 0.28 D WC (cm) -1.69 (0.53) 0.02 0.71 (0.55) 0.20 0.004 -1.33 (0.68) 0.05 0.49 (0.76) 0.52 0.13 D Total fat (g) -430 (506) 0.40 1063 (513) 0.04 0.04 -499 (791) 0.53 100 (853) 0.91 0.79 D Total lean (g) 62.5 (264) 0.81 422 (284) 0.14 0.41 201 (423) 0.64 103 (413) 0.81 0.70 D Total fat mass (%) -0.36 (0.36) 0.33 0.63 (0.35) 0.07 0.05 -0.49 (0.51) 0.34 0.12 (0.50) 0.81 0.60 D Trunk fat (%) -0.37 (0.48) 0.44 0.81 (0.47) 0.09 0.08 -0.51 (0.65) 0.43 0.24 (0.65) 0.71 0.65 Model 2 D WC (cm) -1.67 (0.52) 0.001 0.82 (0.54) 0.13 0.002 -1.31 (0.66) 0.048 0.61 (0.75) 0.42 0.09 D Total fat (g) -428 (508) 0.40 1091 (513) 0.03 0.04 -453 (790) 0.57 146 (858) 0.87 0.77 D Total lean (g) 63.2 (262) 0.81 416 (285) 0.15 0.44 202 (425) 0.63 99.5 (413) 0.81 0.70 D Total fat mass (%) -0.30 (0.36) 0.40 0.63 (0.34) 0.07 0.07 -0.47 (0.51) 0.36 0.15 (0.49) 0.77 0.61 D Trunk fat (%) -0.32 (0.48) 0.51 0.76 (0.46) 0.10 0.10 -0.46 (0.65) 0.48 0.21 (0.64) 0.74 0.75 BMI, Body mass index; WC, Waist circumference b represents changes in outcomes for the increasing number of G allele of the MTNR1B rs10830963 genetic variant Model 1: Adjusted for age, gender, ethnicity and the respective baseline variable Model 2: Model 1 plus BMI at baseline

266 Results

A

Low fat group High fat group 0

-2

-4

-6

-8 ∆ Weight, kg

-10

-12 p=0.004 p=0.29 -14 p interaction=0.01 B

Low fat group High fat group 0

-0.5

-1

2 -1.5

-2 kg/m

-2.5 ∆ BMI, -3

-3.5

-4 p=0.005 p=0.39 -4.5 p interaction=0.02 C

Low fat group High fat group 0

-2

-4

-6

-8

-10 as icmeec, cm ∆ Waist circumference, -12 p=0.001 p=0.13 -14 p interaction=0.002 D

Low fat group High fat group 0

-1

-2

-3

-4

-5

-6 ∆ Total fat, kg -7

-8

-9 p=0.40 p=0.03 -10 p interaction=0.04 Figure 1. Effect of the MTNR1B rs10830963 genetic variant and fat diets on changes in body weight (A), BMI (B), WC (C) and body fat mass (D) at 6 months of diet intervention (black bars, CC genotype; gray bars, CG genotype, white bars, GG genotype). Data are means (SD) after adjusted for age, sex, ethnicity, BMI at the baseline (if appropriate) and value for the respective outcome trait at baseline

267 Chapter 8

Low-fat group High-fat group A

Months Months 0 6 24 0 6 24 0 0

-2 -2

-4 -4

-6 -6

-8 -8 ∆ ∆ Weight,kg ∆ ∆ Weight,kg -10 -10

-12 -12 -14 -14 B

Months Months 0 6 24 0 6 24 0 0

-0.5 -0.5 -1 -1

2 -1.5 2 -1.5 -2 kg/m kg/m -2.5 -2 ∆ BMI, -3 ∆ BMI, -2.5 -3.5 -3 -4 -4.5 -3.5 C Months Months 0 6 24 0 6 24 0 0 -1 -2 -2

-4 -3 -4 -6 -5 -8 -6 -10 -7

∆ ∆ Waist circumference,cm -8 ∆ ∆ Waist circumference,cm -12 -9 -14 D

Months Months 0 6 24 0 6 24 0 0 -1 -1 -2 -3 -2 -4 -3 -5 -6 -4 ∆ Total fat, kg ∆ Total fat, kg -7 -5 -8 -6 -9 -10 -7

Figure 2. Effect of the MTNR1B rs10830963 genetic variant and fat diets on changes in body weight (A), BMI (B), WC (C) and body fat mass (D) at 6 months and 2 years of diet intervention (black circle and solid line, CC genotype; gray circle and gray solid line, CG genotype; white circle and dotted line, GG genotype). Data are means (SE) after adjusted for age, sex, ethnicity, BMI at baseline (if appropriate) and the value for the respective outcome trait at baseline

268 Chapter 6

Supplementary Table 1. Dietary intake and biomarkers of adherence according to the MTNR1B rs10830963 genetic variant and diet group at 6 m and 2 y

At 6 months At 2 years CC CG GG p value CC CG GG p value Low fat diet Dietary intake per day a Energy (kcal) 1640 (613) 1570 (442) 1568 (321) 0.70 1591 (479) 1558 (485) 1528 (445) 0.93 Protein (%) 19.8 (4.4) 19.7 (3.8) 18.8 (3.4) 0.72 20.8 (4.3) 19.7 (3.7) 19.1 (3.6) 0.37 Fat (%) 27.1 (7.7) 24.7 (6.9) 25.2 (7.5) 0.13 27.4 (7.5) 28.1 (8.5) 25.4 (7.0) 0.70 Carbohydrate (%) 54.6 (10.7) 56.1 (9.9) 57.2 (8.5) 0.54 52.1 (9.9) 51.8 (10.1) 53.2 (11.5) 0.94 Biomarkers of adherence Urinary nitrogen (g) b 11.7 (4.2) 11.4 (4.4) 11.7 (4.0) 0.84 11.9 (4.1) 11.8 (4.4) 12.4 (4.7) 0.89 Respiratory quotient c 0.85 (0.04) 0.84 (0.04) 0.85 (0.04) 0.31 0.83 (0.04) 0.84 (0.04) 0.86 (0.04) 0.004 High fat diet Dietary intake per day d Energy (kcal) 1587 (461) 1715 (581) 1491 (250) 0.17 1510 (550) 1377 (403) 1486 (352) 0.57 Protein (%) 20.6 (4.6) 20.3 (5.2) 20.1 (5.0) 0.89 20.4 (5.3) 20.8 (5.4) 19.7 (5.6) 0.86 Fat (%) 34.7 (6.9) 33.9 (7.4) 32.3 (7.3) 0.48 35.8 (8.2) 31.4 (7.1) 31.1 (5.0) 0.04 Carbohydrate (%) 45.3 (7.9) 46.4 (8.6) 47.5 (8.1) 0.55 44.2 (10.3) 48.0 (9.6) 50.3 (5.3) 0.12 Biomarkers of adherence Urinary nitrogen (g) e 11.8 (5.4) 11.5 (4.1) 10.5 (4.4) 0.46 11.7 (4.7) 12.4 (5.1) 12.3 (4.5) 0.58 Respiratory quotient f 0.84 (0.04) 0.83 (0.04) 0.84 (0.05) 0.52 0.83 (0.04) 0.82 (0.04) 0.83 (0.04) 0.80 Data are expressed as mean (SD) a Data were available for 166 individuals at 6 months (CC n=91, CG n=64, GG n=11) and for 88 individuals at 2 years (CC n=39, CG n=42, GG n=7) b Data were available for 266 individuals at 6 months (CC n=140, CG n=101, GG n=25) and for 182 individuals at 2 years (CC n=94, CG n=71, GG n=17) c Data were available for 296 individuals at 6 months (CC n=156, CG n=113, GG n=27) and for 233 individuals at 2 years (CC n=123, CG n=86, GG n=24) d Data were available for 159 individuals at 6 months (CC n=81, CG n=64, GG n=14) and for 81 individuals at 2 years (CC n=49, CG n=23, GG n=9) e Data were available for 260 individuals at 6 months (CC n=137, CG n=101, GG n=22) and for 184 individuals at 2 years (CC n=100, CG n=65, GG n=19) f Data were available for 286 individuals at 6 months (CC n=152, CG n=111, GG n=23) and for 222 individuals at 2 years (CC n=120, CG n=82, GG n=20)

269

CHAPTER 9

Macronutrient-specific effect of the MTNR1B genotype on lipid levels in response to 2-year weight-loss diets

Goni L.1,2, Sun D.3, Heianza Y.3, Wang T.3, Huang T.4, Cuervo M 1,2,5,6, Martínez J.A.1,2,5,6, Shang X.7, Bray G.A.8, Smith S.R.8, Sacks F.M.9, Qi L.3,9,10,11

1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Navarra, Spain 2 Centre for Nutrition Research, University of Navarra, Navarra, Spain 3 Department of Epidemiology, Tulane University, LA, USA 4 Epidemiology Domain, Saw Swee Hoch School of Public Health and Department of Medicine, National University of Singapore, Singapore, Singapore 5 Navarra Institute for Health Research (IdiSNA), Navarra, Spain 6 Biomedical Research Centre Network in Physiopathology of Obesity and Nutrition (CIBERobn), Institute of Health Carlos III, Madrid, Spain 7 Children’s Hospital New Orleans, LA, USA 8 Pennington Biomedical Research Center, Louisiana State University, LA, USA 9 Department of Nutrition, Harvard T.H. Chan School of Public Health, MA, USA 10 Department of Epidemiology, Harvard T.H. Chan School of Public Health, MA, USA 11 Channing Laboratory, Brigham and Women’s Hospital and Harvard Medical School, MA, USA

J Lipid Res Under review (JLR/2017/078634) Impact factor (2016): 4.810 54/286 Biochemistry & Molecular Biology, Q1

Results

ABSTRACT

Compelling evidence indicates that lipid metabolism is into partly control of the circadian system. In this context, it has been reported that a circadian rhythms related Melatonin receptor 1B (MTNR1B) genetic variant influences the dynamics of melatonin secretion, which is involved in the circadian system as a chronobiotic. The objective was to analyze whether the MTNR1B rs10830963 genetic variant was related to changes in lipid levels in response to dietary interventions with different macronutrient distribution in 722 overweight or obese subjects from the POUNDS Lost trial. We did not find a statistically significant association between the MTNR1B genotype and changes in lipid metabolism traits during the 2-year dietary intervention. However, dietary fat intake significantly modified genetic effects on 2- year changes in total cholesterol and LDL cholesterol after adjustment for potential confounders (p interaction=0.006 and 0.001, respectively). In the low-fat diet group, carriers of the sleep disruption G allele (minor allele), showed a greater reduction of total cholesterol (b ± SE=-5.78 ± 2.88 mg/dL, p=0.04) and LDL cholesterol (b ± SE=-7.19 ± 2.37 mg/dL, p=0.003) levels. Conversely, in the high-fat diet group, those subjects carrying the G allele evidenced a smaller decrease in total cholesterol (b ± SE=5.81 ± 2.65 mg/dL, p=0.03) and LDL cholesterol (b ± SE=5.23 ± 2.21 mg/dL, p=0.002) levels. The results indicate that subjects carrying the G allele of the circadian-rhythm related MTNR1B variant may present a bigger impact on total and LDL cholesterol when undertaking an energy restricted low-fat diet.

Keywords Clinical trials, diet and dietary lipids, cholesterol, LDL, genetics

Melatonin receptor 1B, gene-diet interaction, high-fat diet, lipid metabolism, weight-loss intervention

273 Chapter 9

1. INTRODUCTION

There is scientific evidence that lipid metabolism is partly controlled by the circadian system and exhibits differential 24h profiles in major metabolic organs in association with sleep/wake, activity/rest and fast/feeding cycles (1). Plasma lipid concentrations, intestinal absorption and lipid biosynthesis also show a daily rhythmicity as has been reported in different models (2, 3). Moreover, it has been found that disruption of the core circadian clock and peripheral clocks leads to a dysregulation of lipid metabolism (2, 4). Interestingly, Clock mutant mouse exhibited both hyperlipidemia and obesity phenotypes (5).

Melatonin is a hormone secreted mainly by the pineal gland that plays a major role in the regulation of circadian rhythms (6); melatonin treatment has shown beneficial effects on the lipid profile in humans (7). A genetic variant in the melatonin receptor 1B gene (MTNR1B), which encodes one of the two-high affinity receptors of melatonin (8, 9), has been associated with altered melatonin rhythm and melatonin signaling (10, 11). Interestingly, the same genetic variant has also been related to the plasma lipid profile (12, 13). Actually, the MTNR1B rs10830963 polymorphism was related to circulating levels of very low density lipoprotein (VLDL) and triglyceride (12). In addition, the MTNR1B rs10830963 variant was found to interact with dietary fat on lipid levels in an observational study (13). Specifically, Dashti et al. showed that the relation between the MTNR1B rs10830963 genotype and HDL cholesterol was modified by total fat and monounsaturated fatty acids (MUFA) intakes. However, to our knowledge, no study has analyzed the interaction between the MTNR1B rs10830963 polymorphism on long-term changes of lipid metabolism traits in response to dietary interventions. Investigation on such interactions may improve personalized dietary intervention based on the genotype.

The aim of this study was to examine potential interactions between the MTNR1B rs10830963 genotype and weight-loss diets varying in fat content on changes in lipid metabolism traits during 2-years of dietary intervention within the POUNDS Lost trial.

2. MATERIALS AND METHODS

2.1. Study participants

The POUNDS Lost trial is a 2-year randomized clinical trial (clinical trial reg. no. NCT00072995) designed to compare the effects of four energy-reduced diets with different macronutrient composition on weight loss. The study was conducted at two sites: Harvard School of Public Health and Brigham and Women’s Hospital in Boston, MA, and the Pennington Biomedical

274 Results

Research Center of Louisiana State University System, Baton Rouge, LA; from October 2004 through December 2007. The study design and methods have been previously described in detail (13). Briefly, 811 overweight or obese (body mass index (BMI) 25-40 kg/m2) participants were randomly assigned to one of four energy-reduced diets during a 2 year follow-up time. The target percentages of energy derived from fat, protein and carbohydrate in the four diets were: 20, 15 and 60%; 20, 25 and 55%; 40, 15 and 45%; and 40, 25 and 35%, respectively. In this two-by-two factorial design two diets were low fat (20%), two diets were high fat (40%), two diets were average in protein (15%) and two diets were high in protein (25%). Major exclusion criteria were the presence of diabetes or unstable cardiovascular disease, the use of medications that affect body weight, and insufficient motivation (13). The study was approved by the human subjects committee at each institution and by a data and safety monitoring board appointed by the National Heart, Lung and Blood Institute. All participants gave written informed consent.

2.2. Measurements

Body weight was measured in the morning before breakfast at baseline, 6 months and 2 years of follow-up. Height was measured at baseline. BMI was calculated as weight (kg) / height (m2). In the present study ethnicity was self-reported and grouped as white, black and others. Fasting blood samples were obtained at routine times in clinical settings at baseline, 6 months and 2 years. Levels of serum lipids (TG, total cholesterol and HDL cholesterol) were analyzed in the clinical Laboratory at Pennington using the Synchron CX7 (Beckman Coulter). LDL cholesterol was obtained for each participant according to the following equation: total cholesterol – HDL cholesterol – TG/5 (15). However, when TG concentration was >400 mg/dL LDL cholesterol was measured directly by Synchron CX7 (Beckman Coulter). Dietary intake was assessed in a random sample of 50% of the participants by a review of the 5 days diet record at baseline and by 24-h recall during a telephone interview on 3 nonconsecutive days at 6 months and 2 years, in order to assess the nutritional adherence across the intervention.

2.3. Genotyping

DNA was extracted from the buffy coat fraction of centrifuged blood by the QIAmp Blood Kit (Qiagen). Genotyping of the previously reported MTNR1B rs10830963 genetic variant was successfully done in 722 of 811 total participants using the Open Array SNP Genotyping System (Biotrove) (16). The genotyping success rate was near 90%. The genotype frequency was CC

275 Chapter 9

54.4%, CG 37.8% and GG 7.8%; being the frequency of the minor allele (G allele) 0.27 in the study population. The genotype distribution was in Hardy-Weinberg equilibrium (HWE) in both total population and major ethnic group (whites) (p > 0.05).

2.4. Statistical analysis

In the present study, 722 subjects with baseline genotyped data of the MTNR1B rs10830963 variant were included. Among them, 528 participants completed the intervention after 2 years. The primary outcome of the current study was the change in lipid levels (TG, total cholesterol, LDL cholesterol and HDL cholesterol) during the 2 years of intervention. To compare baseline characteristics across genotypes chi-squared test for categorical variables and general linear models for continuous variables were used. General linear models (model 1 adjusted for age, gender, ethnicity, baseline BMI, the respective baseline variable and lipid- lowering medication used; model 2 adjusted for model 1 plus weight loss) were performed for comparison of changes from baseline in lipid levels across genotype groups according to low- or high-fat group at 6 months and 2 years intervention. The interaction term (e.g., MTNR1B genotype × high-/low-fat diet group) was included in the models to test gene-diet intervention interactions. Additive genetic models were used in the analysis. Statistical analyses were performed using STATA/SE version 12.0 (StataCorp, College Station, TX, USA). The statistical significance was considered for p value < 0.05. Moreover, we used Bonferroni correction to adjust p values for 4 independent test (triglycerides, total cholesterol, LDL cholesterol and HDL cholesterol). Thus, a p value <0.012 was considered statistically significant after adjustment of multiple comparisons.

3. RESULTS

Baseline characteristics of the participants according to the MNTR1B rs10830963 genotype are presented in Table 1. The distribution of the polymorphism was similar by gender and diet groups; meanwhile statistically significant differences were observed by ethnicity (p<0.001). No associations of the MNTR1B genotype with baseline lipid levels were observed. Changes in lipid levels at 6 months and 2 years of follow-up were not related to the MTNR1B genotype after adjustment for age, sex, ethnicity, BMI at baseline, baseline value for the respective outcome, lipid-lowering medication used and diet group (data not shown).

Dietary fat (high- versus low-fat intake) showed significantly differential effects on changes in total cholesterol and LDL cholesterol at 2 years of the intervention, depending on the MTNR1B

276 Results rs10830963 genetic variant, after adjusting for age, gender, ethnicity, BMI at baseline, the respective baseline variable and lipid-lowering medication use (p interaction=0.006 and 0.001, respectively) (Table 2 and Figures 1 and 2). These genotype-diet interactions were statistically significant after correction for multiple testing (p < 0.012 based on Bonferroni correction for 4 test). Within the high-fat diet group, an increasing number of the G allele was associated with greater decreases in total cholesterol and LDL cholesterol (p=0.045 and 0.003, respectively) (b ± SE=-5.78 ± 2.88 mg/dL per G allele, p=0.04) and LDL cholesterol (b ± SE=-7.19 ± 2.37 mg/dL per G allele, p=0.003). Meanwhile, an opposite effect was observed among participants in the high-fat diet group. Thus, carriers of the G allele showed a positive association with increases in total cholesterol and LDL cholesterol (p=0.029 and 0.019, respectively). (b ± SE=5.81 ± 2.65 mg/dL per G allele, p=0.03) and LDL cholesterol (b ± SE=5.23 ± 2.21 mg/dL per G allele, p=0.02).

In order to detect whether the observed effects were mediated by weight loss, the analyzed models were further adjusted for body weight loss (model 2) (Table 2). We found that the gene-diet interactions on changes in total cholesterol and LDL cholesterol remained significant (p for interaction=0.008 and 0.001, respectively). Only in the low-fat diet group was the association between total cholesterol and the genotype attenuated (p=0.063). When we analyzed the data among white individuals similar results were found (p interaction 0.025 and 0.003 in model 2 for total cholesterol and LDL cholesterol, respectively) (Supplemental Table 1).

There were no significant interactions between the genotype and the low-/high-fat diet on changes in HDL cholesterol and TG at 2 years of follow-up (both p for interaction >0.05). In addition, the MTNR1B rs10830963 genetic variant did not interact with dietary protein intake on changes in lipid levels (data not shown).

4. DISCUSSION

In one of the largest available randomized dietary intervention trials on weight loss, we, for the first time, report a significant interaction between the circadian rhythm related MTNR1B rs10830963 genetic variant and dietary fat intake on changes in total cholesterol and LDL cholesterol. Our results indicate that an increasing number of the G allele was associated with greater decreases in total cholesterol and LDL cholesterol in response to the low-fat diet, whereas an opposite effect was found in the high-fat diet group.

277 Chapter 9

Compelling evidence has shown that the circadian system plays an important role in coordinating lipid metabolic pathways through rhythmic activation or repression of genes involved in lipid metabolism, either directly or in directly by controlling other transcription factors (17). In addition, the disruption of the core molecular clock results in abnormal lipid metabolism, including altered fat storage and lipid transport, and deficits in absorption of dietary lipids (1, 2). In this sense, melatonin is one of the chronobiotics used by the central master clock to synchronize circadian rhythms (6). Regarding lipid metabolism, it has been demonstrated that melatonin treatment can improve dyslipidemia in both, animal and human studies (7). In humans, daily administration of melatonin for 2 months significantly improved LDL cholesterol among subjects with features of the Metabolic Syndrome (18). Moreover, treatment with melatonin and zinc decreased the levels of TG, total cholesterol, LDL cholesterol and increased the levels of HDL cholesterol in type 2 diabetic patients poorly controlled with metformin (19). Despite the fact that the effect of melatonin on lipid profiles has been widely studied, the mechanisms by which the MTNR1B rs10830963 affects lipid metabolism remains unknown. However, the MTNR1B genetic variant has been recently associated with melatonin levels and melatonin signaling (10, 11). On the one hand, Lane et al. reported that rs10830963 G allele carriers showed a disruption of melatonin rhythm, since subjects presented a later melatonin offset and a longer duration of elevated melatonin levels (10). Given that melatonin appears to be involved in various lipid phenotypes, it can be speculated that the effect of the MTNR1B genetic variant on dynamics of melatonin expression thereby could influence lipid levels. On the other hand, Tuomi et al. reported that subjects carrying 1 or 2 MTNR1B rs10830963 G alleles showed a 2- and 4-fold increase in MTNR1B mRNA expression in human pancreatic islets, respectively, compared with subjects carrying the CC genotype (11). Although the results by Tuomi et al. did not show a direct impact on lipid levels, such findings suggest that MTNR1B rs10830963 might affect MTNR1B mRNA expression in other cell types related to lipid metabolism.

Interestingly, we found that dietary fat intake modified the effect of the MTNR1B rs10830963 genetic variant on changes in total cholesterol and LDL cholesterol, which provide suggestive implications in preventive medicine and clinical practice. In this context, a meta-analysis found nominal significant interactions between the MTNR1B genotype and fat intake (total fat and MUFA) on HDL cholesterol levels (13). However, the results did not pass the prespecified Bonferroni-corrected significance level. Although the mechanisms underlying the observed MTNR1B rs10830963 gene-dietary fat interaction are unknown, there is evidence that a high- fat diet could alter the expression and the rhythmic mRNA expression levels of circadian-clock

278 Results genes and circadian clock-controlled lipogenic genes (20–22). For example, Sun L. et al. observed a rhythmic expression of the clock-controlled output gene Ppar-a and downstream lipid metabolism genes (Srebp-1c, Fas and Acc1) in normally fed mice (22). Meanwhile, when mice were fed with a high-fat diet, the rhythmic expression in the liver of such genes was significantly altered. Moreover, the effects of the genotype on changes in lipid metabolism traits showed opposite trends in participants with low v.s high fat intake. The results are in line with the “differential susceptibility hypothesis” which proposes that vulnerability genes or risk alleles may function like plasticity genes since genetic risk can be modified by environmental exposures, including dietary factors (23–26). In other words, some individuals might be more responsive to environmental influences in a “for-better-and-for-worse” manner because of the genetic background (24). Consistent with this hypothesis, we observed that carriers of the risk allele might function as either a protective or a detrimental factor, depending on the differences in dietary fat intake. In the present study, we failed to ascertain a genotype-diet interaction on changes in triglycerides and HDL cholesterol. The reason for this differential genetic effect on various lipid phenotypes is unclear. However, it can be hypothesized that different mechanisms might drive the reported gene-diet interactions for changes on different lipid components.

Our data indicate that the gene-diet interaction on changes in total cholesterol and LDL cholesterol remained statistically significant after additional adjustment for weight loss, although a slight attenuation of the effect was observed on changes in total cholesterol. These results suggest that the effects on changes in total cholesterol and LDL cholesterol concentrations might be independent of weight loss, although weight reduction has been shown to induce beneficial effects on lipid profile, mainly on total cholesterol and LDL cholesterol (27). However, in our study weight loss was correlated with TG levels and HDL cholesterol but not with total cholesterol and LDL cholesterol at 2 years of the intervention (28).

To the best of our knowledge, this is the first study to analyze interactions between a MTNR1B genetic variant and dietary fat intake on changes in plasma lipids in a large and long-term dietary intervention trial. These findings add novel insights into the role of the circadian rhythms related MTNR1B genetic variant on metabolic responses (16). However, the present study has some limitations that should be addressed. We did not measure circulating melatonin levels in the study population, which prevented the potential analysis of the relationship between the genetic variant and circulating melatonin levels. Nonetheless, according to the Mendelian randomization principle, a genetic variant could be a surrogate for

279 Chapter 9 the biomarker in causal inference, because it is less likely to be affected by confounding and reverse causation (29). In addition, it is difficult to determine which macronutrient plays the key role of the observed interactions because the low-fat intake is characterized by high- carbohydrate intake, and vice versa, to maintain energy balance. Finally, the results should be replicated in order to be extended to other ethnic groups, because most of the participants were whites (around 80%), and to rule out the possibility of false positive findings.

In summary, these results suggest that carriers of the G allele of the circadian rhythms related MTNR1B rs10830963 genetic variant may benefit more in the improvement of their lipid profile by choosing a low-fat diet instead of a high-fat diet. These findings may lend support to personalized dietary interventions in improvement of lipid metabolism.

Acknowledgments

The authors thank all participants of the study for their dedication and contribution to the research.

Grant support

The study is supported by grants from the National Heart, Lung, and Blood Institute (HL- 071981, HL-034594, HL-126024), the National Institute of Diabetes and Digestive and Kidney Diseases (DK-091718, DK-100383, DK-078616), the Boston Obesity Nutrition Research Center (DK-46200), and United States-Israel Binational Science Foundation Grant 2011036. L.G. is a recipient of a pre-doctoral and a mobility grant from the Spanish Ministry of Education, Culture and Sport. Y.H. is a recipient of a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science. L.Q. was a recipient of the American Heart Association Scientist Development Award (0730094N).

280 Results

REFERENCES 1. Kumar Jha, P., E. Challet, and A. Kalsbeek. 2015. Circadian rhythms in glucose and lipid metabolism in nocturnal and diurnal mammals. Mol. Cell. Endocrinol. 418: 74–88. 2. Gooley, J. J. 2016. Circadian regulation of lipid metabolism. Proc. Nutr. Soc. 75: 440–450. 3. Bailey, S. M., U. S. Udoh, and M. E. Young. 2014. Circadian regulation of metabolism. J. Endocrinol. 222: R75–R96. 4. Obayashi, K., K. Saeki, J. Iwamoto, N. Okamoto, K. Tomicka, S. Nezu, Y. Ikada, and N. Kurumatani. 2013. Exposure to light at night, nocturnal urinary melatonin secretion, and obesity/dyslipidemia in the elderly: a cross-sectional analysis of the HEIJO-KYO study. J. Clin. Endocrinol. Metab. 98: 337–344. 5. Turek, F. W., C. Joshu, A. Kohsaka, E. Lin, G. Ivanova, E. McDearmon, A. Laposky, S. Losee-Olson, A. Easton, D. R. Jensen, R. H. Eckel, J. S. Takahashi, and J. Bass. 2005. Obesity and metabolic syndrome in circadian Clock mutant mice. Science (80-. ). 308: 1043–1045. 6. Cipolla-Neto, J., F. G. Amaral, S. C. Afeche, D. X. Tan, and R. J. Reiter. 2014. Melatonin, energy metabolism, and obesity: A review. J. Pineal Res. 56: 371–381. 7. Sun, H., F. Huang, and S. Qu. 2015. Melatonin: a potential intervention for hepatic steatosis. Lipids Health Dis. 14: 75–80. 8. Dubocovich, M. L., P. Delagrange, D. N. Krause, D. Sugden, D. P. Cardinali, and J. Olcese. 2010. Nomenclature, classification, and pharmacology of G Protein-Coupled Melatonin Receptors. Pharmocological Rev. 62: 343–380. 9. Gamble, K. L., R. Berry, S. J. Frank, and M. E. Young. 2014. Circadian clock control of endocrine factors. Nat. Rev. Endocrinol. 10: 466–475. 10. Lane, J. M., A. M. Chang, A. C. Bjonnes, D. Aeschbach, C. Anderson, B. E. Cade, S. W. Cain, C. A. Czeisler, S. A. Gharib, J. J. Gooley, D. J. Gottlieb, S. F. A. Grant, E. B. Klerman, D. S. Lauderdale, S. W. Lockley, M. Munch, S. Patel, N. M. Punjabi, S. M. W. Rajaratnam, M. Rueger, M. A. St. Hilaire, N. Santhi, K. Scheuermaier, E. Van Reen, P. C. Zee, S. A. Shea, J. F. Duffy, O. M. Buxton, S. Redline, F. A. J. L. Scheer, and R. Saxena. 2016. Impact of common diabetes risk variant in MTNR1B on sleep, circadian, and melatonin physiology. Diabetes. 65: 1741–1751. 11. Tuomi, T., C. L. F. Nagorny, P. Singh, H. Bennet, Q. Yu, I. Alenkvist, B. Isomaa, B. Östman, J. Söderström, A. K. Pesonen, S. Martikainen, K. Räikkönen, T. Forsén, L. Hakaste, P. Almgren, P. Storm, O. Asplund, L. Shcherbina, M. Fex, J. Fadista, A. Tengholm, N. Wierup, L. Groop, and H. Mulder. 2016. Increased melatonin signaling is a risk factor for type 2 diabetes. Cell Metab. 23: 1067–1077. 12. DeMenna, J., S. Puppala, G. Chittoor, J. Schneider, J. Y. Kim, G. Q. Shaibi, L. J. Mandarino, R. Duggirala, and D. K. Coletta. 2014. Association of common genetic variants with diabetes and metabolic syndrome related traits in the Arizona insulin resistance registry: A focus on Mexican American families in the southwest. Hum. Hered. 78: 47–58. 13. Dashti, H. S., J. L. Follis, C. E. Smith, T. Tanaka, M. Garaulet, D. J. Gottlieb, A. Hruby, P. F. Jacques, J. C. Kiefte-De Jong, S. Lamon-Fava, F. A. J. L. Scheer, T. M. Bartz, L. Kovanen, M. K. Wojczynski, A. C. Frazier-Wood, T. S. Ahluwalia, M. M. Perälä, A. Jonsson, T. Muka, I. P. Kalafati, V. Mikkilä, and J. M. Ordovás. 2015. Gene-environment interactions of circadian-related genes for cardiometabolic traits. Diabetes Care. 38: 1456–1466. 14. Sacks, F. M., G. A. Bray, V. J. Carey, S. R. Smith, D. H. Ryan, S. D. Anton, C. McManus, C. M. Champagne, L. M. Bishop, N. Laranjo, M. S. Leboff, J. C. Rood, L. de Jonge, F. L. Greenway, C. M. Loria, E. Obarzanek, and D. A. Williamson. 2009. Comparison of weight-loss diets with different compositions of fat, protein, and carbohydrates. N Engl J Med. 360: 859–873. 15. Friedewald, W. T., R. I. Levy, and D. S. Fredrickson. 1972. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin. Chem. 18: 499–502. 16. Mirzaei, K., M. Xu, Q. Qi, L. De Jonge, G. A. Bray, F. Sacks, and L. Qi. 2014. Variants in glucose- and circadian

281 Chapter 9 rhythm-related genes affect the response of energy expenditure to weight-loss diets: the POUNDS LOST Trial. Am. J. Clin. Nutr. 99: 392–399. 17. Gooley, J. J., and E. C. hern P. Chua. 2014. Diurnal regulation of lipid metabolism and applications of circadian lipidomics. J. Genet. genomics. 231–250. 18. Koziróg, M., A. R. Poliwczak, P. Duchnowicz, M. Koter-Michalak, J. Sikora, and M. Broncel. 2011. Melatonin treatment improves blood pressure, lipid profile, and parameters of oxidative stress in patients with metabolic syndrome. J. Pineal Res. 50: 261–266. 19. Kadhim, H. M., S. H. Ismail, K. I. Hussein, I. H. Bakir, A. S. Sahib, B. H. Khalaf, and S. A. R. Hussain. 2006. Effects of melatonin and zinc on lipid profile and renal function in type 2 diabetic patients poorly controlled with metformin. J. Pineal Res. 41: 189–193. 20. Yanagihara, H., H. Ando, Y. Hayashi, Y. Obi, and A. Fujimura. 2006. High-fat feeding exerts minimal effects on rhythmic mRNA expression of clock genes in mouse peripheral tissues. Chronobiol. Int. 23: 905–914. 21. Hsieh, M.-C., S.-C. Yang, H.-L. Tseng, L.-L. Hwang, C.-T. Chen, and K.-R. Shieh. 2010. Abnormal expressions of circadian-clock and circadian clock-controlled genes in the livers and kidneys of long-term, high-fat-diet-treated mice. Int. J. Obes. (Lond). 34: 227–239. 22. Sun, L., Y. Wang, Y. Song, X. R. Cheng, S. Xia, M. R. T. Rahman, Y. Shi, and G. Le. 2015. Resveratrol restores the circadian rhythmic disorder of lipid metabolism induced by high-fat diet in mice. Biochem. Biophys. Res. Commun. 458: 86–91. 23. Belsky, J., C. Jonassaint, M. Pluess, M. Stanton, B. Brummett, and R. Williams. 2009. Vulnerability genes or plasticity genes? Mol. Psychiatry. 14: 746–754. 24. Belsky, J., and S. Hartman. 2014. Gene-environment interaction in evolutionary perspective: differential susceptibility to environmental influences. World Psychiatry. 13: 87–89. 25. Hartman, S., and J. Belsky. 2016. An evolutionary perspective on family studies: differential susceptibility to environmental influences. Fam. Process. 55: 1–13. 26. Dalle Molle, R., H. Fatemi, A. Dagher, R. D. Levitan, P. P. Silveira, and L. Dubé. 2016. Gene and environment interaction: is the differential susceptibility hypothesis relevant for obesity? Neurosci. Biobehav. Rev. 73: 326–339. 27. Poobalan, A., L. Aucott, W. C. S. Smith, A. Avenell, R. Jung, J. Broom, and A. Grant. 2004. Effects of weight loss in overweight/obese individuals and long-term lipids outcomes a systematic review. Obes. Rev. 5: 43–50. 28. Zhang, X., Q. Qi, G. A. Bray, F. B. Hu, F. M. Sacks, and L. Qi. 2012. APOA5 genotype modulates 2-y changes in lipid profile in response to weight-loss diet intervention: The Pounds Lost Trial. Am. J. Clin. Nutr. 96: 917–922. 29. Qi, L. 2009. Mendelian randomization in nutritional epidemiology. Nutr. Rev. 67: 439–450.

282 Results

Table 1. Baseline characteristics of the study participants according to MTNR1B rs10830963 genotypes

CC CG GG p value (n=393) (n=273) (n=56) Age (years) 50.8 (9.2) 51.0 (9.1) 53.0 (10.3) 0.24 Sex 0.72 Male 149 (37.9) 112 (41.0) 22 (39.3) Female 244 (62.1) 161 (59.0) 34 (60.7) Race or ethnic group <0.001 White 279 (71.0) 247 (90.5) 49 (87.5) Black 96 (24.4) 12 (4.4) 3 (5.4) Hispanic or other 18 (4.6) 14 (5.1) 4 (7.1) Diet group 0.69 Low fat 198 (50.4) 134 (49.1) 31 (55.4) High fat 195 (49.6) 139 (50.2) 25 (44.6) TG (mg/dL) 137.4 (86.8) 151.4 (85.4) 151.5 (84.4) 0.09 Total cholesterol (mg/dL) 200.0 (37.3) 204.4 (37.2) 206.2 (33.9) 0.22 LDL cholesterol (mg(dL) 125.0 (32.0) 125.9 (32.7) 126.4 (31.2) 0.92 HDL cholesterol (mg/dL) 48.2 (13.3) 48.9 (13.8) 51.2 (20.3) 0.31 Body weight, kg 93.7 (15.0) 92.8 (16.0) 92.0 (17.5) 0.66 Body weight loss at 2 years, kg -3.7 (0.4) -4.1 (0.5) -6.0 (1.0) 0.13 a BMI, kg/m2 32.9 (3.8) 32.3 (3.9) 32.5 (3.7) 0.11 BMI loss at 2 years, kg/m2 -1.3 (0.1) -1.4 (0.2) -2.1 (0.3) 0.12 a Data were calculated by the c2 test for categorical variables and ANOVA for continuous variables. Data are expressed as n (%) or mean (SD) a Adjusted for age, sex, ethnicity, baseline value for the respective outcome and diet group

283 Chapter 9

Table 2. Effect of the MTNR1B rs10830963 genetic variant on changes in lipid metabolism traits in response to a low-/high-fat diet at 2 years of diet intervention

Low fat High fat

(n=270) (n=258) p interaction b (SE) p b (SE) p Model 1 D TG, mg/dL 5.37 (5.51) 0.33 0.65 (5.44) 0.90 0.65 D Total cholesterol, mg/dL -5.78 (2.88) 0.04 5.81 (2.65) 0.03 0.006 D LDL cholesterol, mg/dL -7.19 (2.37) 0.003 5.23 (2.21) 0.02 0.001 D HDL cholesterol, mg/dL 0.71 (0.67) 0.29 0.70 (0.74) 0.34 0.90 Model 2 D TG, mg/dL 8.62 (5.30) 0.10 0.51 (5.18) 0.92 0.44 D Total cholesterol, mg/dL -5.41 (2.90) 0.06 5.81 (2.66) 0.03 0.008 D LDL cholesterol, mg/dL -6.85 (2.38) 0.004 5.23 (2.21) 0.02 0.001 D HDL cholesterol, mg/dL 0.23 (0.63) 0.71 0.74 (0.69) 0.28 0.76 Data were calculated by using linear regression models. The interaction term was included in the models to test gene-diet interactions. b represents changes in outcomes for the increasing number of G allele of the rs10830963 variant Model 1: Adjusted for age, gender, ethnicity, BMI at baseline, the respective baseline variable and lipid-lowering medication use Model 2: Adjusted for age, gender, ethnicity, BMI at baseline, the respective baseline variable, lipid-lowering medication use and body weight loss at each intervention time

284 Results

A

Low-fat group High-fat group 20 15 10 5 CC 0 CG -5 GG -10 -15

oa hlseo (mg/dL) ∆ Total cholesterol -20 -25 p=0.06 p=0.03 -30 p interaction=0.008

B

Low-fat group High-fat group 15

10

5

0

(mg/dL) CC

-5 CG

-10 GG holesterol c -15

∆ LDL -20

-25 p=0.004 p=0.02 -30 p interaction=0.001

Figure 1. Interaction between the MTNR1B rs10830963 genetic variant and dietary fat intervention on changes in total cholesterol (A) and LDL cholesterol (B) at 2 years of diet intervention. Data are means (SE) after adjusted for age, sex, ethnicity, BMI at baseline, the value for the respective outcome trait at baseline and lipid-lowering medication use. Low-fat group sample sizes: CC n=146, CG n=97 and GG n=27. High-fat group sample sizes: CC n=137, CG n=100 and GG n=21.

285 Chapter 9

Low-fat group High-fat group A

15 15

5 5

-5 -5

-15 -15 oa hlseo (mg/dL) ∆ Total cholesterol

-25 (mg/dL) ∆ Total cholesterol -25

-35 0 6 24 -35 Months 0 6 24 Months B

15 15

5 5

-5 c (mg/dL) -

c (mg/dL) -5 - ∆ LDL ∆ LDL -15 -15

-25 0 6 24 -25 Months 0 6 24 Months

Figure 2. Effect of the MTNR1B rs10830963 genetic variant and fat diets on changes in total cholesterol (A) and LDL cholesterol (B) at 6 months and 2 years of diet intervention (black circle and solid line, CC genotype; gray circle and gray solid line, CG genotype; white circle and dotted line, GG genotype). Data are means (SE) after adjusted for age, sex, ethnicity, BMI at baseline, the value for the respective outcome trait at baseline and lipid-lowering medication use. At 6 months low-fat group sample sizes: CC n=167, CG n=121 and GG n=27. At 6 months high-fat group sample sizes: CC n=165, CG n=115 and GG n=23.At 24 months low-fat group sample sizes: CC n=146, CG n=97 and GG n=27. At 2 years high-fat group sample sizes: CC n=137, CG n=100 and GG n=21.

286 Results

Supplemental Table 1. Effect of the MTNR1B rs10830963 genetic variant on changes in lipid metabolism traits in response to a low-/high-fat diet at 2 years of diet intervention among white subjects

Low fat (n=240) High fat (n=235) p interaction b (SE) p b (SE) p Model 1 D Triglycerides, mg/dL 7.81 (6.18) 0.21 1.12 (5.93) 0.85 0.51 D Total cholesterol, mg/dL -4.70 (2.77) 0.09 4.24 (2.81) 0.13 0.02 D LDL cholesterol, mg/dL -6.59 (2.25) 0.004 3.51 (2.30) 0.13 0.002 D HDL cholesterol, mg/dL 0.87 (0.72) 0.23 0.30 (0.79) 0.70 0.55 Model 2 D Triglycerides, mg/dL 11.1 (5.91) 0.06 0.19 (5.69) 0.97 0.25 D Total cholesterol, mg/dL -4.15 (2.78) 0.14 4.16 (2.81) 0.14 0.02 D LDL cholesterol, mg/dL -6.13 (2.26) 0.007 3.44 (2.31) 0.14 0.003 D HDL cholesterol, mg/dL 0.44 (0.68) 0.52 0.45 (0.74) 0.54 0.98 Data were calculated by using linear regression models. The interaction term was included in the models to test gene-diet interactions. b represents changes in outcomes for the increasing number of G allele of the rs10830963 variant Model 1: Adjusted for age, gender, ethnicity, BMI at baseline, the respective baseline variable and lipid-lowering medication use Model 2: Adjusted for age, gender, ethnicity, BMI at baseline, the respective baseline variable, lipid-lowering medication use and body weight loss at each intervention time

287

CHAPTER 10

Effect of the interaction between diet composition and the PPM1K genetic variant on insulin resistance and ß cell function markers during weight loss: results from the NUGENOB randomized trial

Goni L.1,2, Qi L.3,9,10,11, Cuervo M.1,2,5,6, Milagro F.I.1,2,5,6, Saris W.H.7, MacDonald I.A.8, Langin D.9,10, Astrup A.11, Arner P.12, Oppert J.M.13, Svendstrup M.14,15, Blaak E.E.7, Sørensen T.I.A.15,16,17, Hansen T.15, Martínez J.A.1,2,5,6

1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Navarra, Spain 2 Centre for Nutrition Research, University of Navarra, Navarra, Spain 3 Department of Epidemiology, Tulane University, LA, USA 4 Department of Nutrition, Harvard T.H. Chan School of Public Health, MA, USA 10 Department of Epidemiology, Harvard T.H. Chan School of Public Health, MA, USA 11 Channing Laboratory, Brigham and Women’s Hospital and Harvard Medical School, MA, USA 5 Navarra Institute for Health Research (IdiSNA), Navarra, Spain 6 Biomedical Research Centre Network in Physiopathology of Obesity and Nutrition (CIBERobn), Institute of Health Carlos III, Madrid, Spain 7 Department of Human Biology, Maastricht University Medical Center +, Maastricht, The Netherlands 8 School of Life Sciences, University of Nottingham, Nottingham, UK 9 Institut National de la Santé et de la Recherche Médicale, Institute of Metabolic and Cardiovascular Diseases, Toulouse, France 10 University of Toulouse, Institute of Metabolic and Cardiovascular Diseases, Paul Sabatier University, Toulouse, France 11 Department of Nutrition, Exercise and Sports, University of Copenhagen, Copenhagen, Denmark 12 Department of Medicine Huddinge, Karolinska Institute, Stockholm, Sweden 13 Department of Nutrition, Pitie-Salpetriere University Hospital, University Pierre et Marie-Curie-Paris 6, Institute of Cardiometabolism and Nutrition, Center for Research on Human Nutrition, Ile-de-France, Paris, France 14 Danish Diabetes Academy, Odense, Denmark 15 Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen, Denmark 16 Department of Public Health, University of Copenhagen, Denmark 17 Department of Clinical Epidemiology, Bispebjerg and Frederiksberg Hospitals, Copenhagen, Denmark

Am J Clin Nutr. 2017 DOI 10.3945/ajcn.117.156281 Impact factor (2016): 6.929 3/81 Nutrition & Dietetics, Q1

Goni L, Qi L, Cuervo M, Milagro FI, Saris WH, MacDonald IA, Langin D, Astrup A, Arner P, Oppert JM, Svendstrup M, Blaak EE, Sorensen TIA, Hansen T, Martinez JA. Effect of the interaction between diet composition and the PPM1K genetic variant on insulin resistance and ß cell function markers during weight loss: results from the NUGENOB randomized trial. American Journal of Clinical Nutrition, 2017, 106(3):902- 908. https://doi.org/10.3945/ajcn.117.156281

CHAPTER 11

Influence of fat intake and BMI on the association of rs1799983 NOS3 polymorphism with blood pressure levels in an Iberian population

Goni L.1,2, Cuervo M.1,2,3,4, Milagro F.I.1,2,3,4, Martínez J.A.1,2,3,4

1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Navarra, Spain 2 Centre for Nutrition Research, University of Navarra, Navarra, Spain 3 Navarra Institute for Health Research (IdiSNA), Navarra, Spain 4 Biomedical Research Centre Network in Physiopathology of Obesity and Nutrition (CIBERobn), Institute of Health Carlos III, Madrid, Spain

Eur J Nutr. 2017 DOI 10.1007/s00394-016-1203-3 Impact factor (2016): 4.370 14/81 Nutrition & Dietetics, Q1

Results

Eur J Nutr (2017) 56:1589–1596 DOI 10.1007/s00394-016-1203-3

ORIGINAL CONTRIBUTION

Influence of fat intake and BMI on the association of rs1799983 NOS3 polymorphism with blood pressure levels in an Iberian population

Leticia Goni1,2 · Marta Cuervo1,2,3,4 · Fermín I. Milagro1,2,3 · J. Alfredo Martínez1,2,3,4

Received: 27 May 2015 / Accepted: 11 March 2016 / Published online: 19 March 2016 © Springer-Verlag Berlin Heidelberg 2016

Abstract individuals, T allele carriers showed higher DBP than GG Purpose There is controversy about the effect of the genotype. rs1799983 nitric oxide synthase (NOS3) genetic variant Conclusion The present study evidenced that rs1799983 on hypertension and blood pressure (BP) levels. The aims NOS3 polymorphism could be associated with hyperten- of the current study were to examine whether rs1799983 sion and DBP among Southern Europeans, being this asso- affects BP levels and to identify potential interactions ciation influenced by dietary fat (SFA and MUFA) and between this polymorphism and other non-genetic risk body mass index. factors. Methods A total of 705 subjects were examined for Keywords Hypertension · Blood pressure · NOS3 · anthropometric and body composition measurements, BP, Saturated fatty acids · Monounsaturated fatty acids · dietary habits and physical activity. Oral epithelial cells Obesity were collected for the identification of rs1799983 using Luminex® 100/200TM System. Results After adjusted for covariates, TT genotype Introduction showed a 2.30-fold higher predisposition of hypertension than GG genotype subjects. According to BP levels, for Hypertension contributes to the burden of heart disease, each risk allele diastolic blood pressure (DBP) increased stroke, kidney failure and premature mortality [1]. In fact, in 1.99 mmHg. Significant interactions between rs1799983 hypertension complications account for 9.4 million deaths and saturated fatty acids (SFA) and monounsaturated fatty worldwide every year [2]. Key risk factors include age, acids (MUFA) were found. Moreover, an interaction with race, endocrine and metabolic disorders, lifestyle behaviors body weight status was observed. Among overweight and genetics, among others [1]. In this context, familial and twin studies have estimated the heritable component of blood pressure (BP) to be about 30–60 % [3]. Moreover, genomewide association studies * J. Alfredo Martínez (GWAS) have identified a large number of polymorphisms [email protected] associated with BP or hypertension, which are located in or near genes involved in the renin–angiotensin–aldosterone 1 Department of Nutrition, Food Sciences and Physiology, University of Navarra, Irunlarrea, 1, 31008 Pamplona, system, related to enzymes and receptors of the mineral Navarra, Spain and glucocorticoid pathways and associated with proteins 2 Centre for Nutrition Research, University of Navarra, implicated in the structure and/or regulation of vascular Pamplona, Spain tone [4]. 3 CIBER Fisiopatología de la Obesidad y Nutrición Among them, the nitric oxide synthase (NOS3) gene is (CIBERobn), Instituto de Salud Carlos III, Madrid, Spain regarded as one of the putative candidate genes for BP reg- 4 Instituto de Investigación Sanitaria de Navarra (IdiSNA), ulation and hypertension, since it is involved in the produc- Pamplona, Spain tion of nitric oxide (NO) which has vasodilator effects (i.e.,

1 3

313 Chapter 11

ABSTRACT

Purpose There is controversy about the effect of the rs1799983 nitric oxide synthase (NOS3) genetic variant on hypertension and blood pressure (BP) levels. The aims of the current study were to examine whether rs1799983 affects BP levels and to identify potential interactions between this polymorphism and other non-genetic risk factors.

Methods A total of 705 subjects were examined for anthropometric and body composition measurements, BP, dietary habits and physical activity. Oral epithelial cells were collected for the identification of rs1799983 using Luminex® 100/200TM System.

Results After adjusted for covariates, TT genotype showed a 2.30-fold higher predisposition of hypertension than GG genotype subjects. According to BP levels, for each risk allele diastolic blood pressure (DBP) increased in 1.99 mmHg. Significant interactions between rs1799983 and saturated fatty acids (SFA) and monounsaturated fatty acids (MUFA) were found. Moreover, an interaction with body weight status was observed. Among overweight individuals, T allele carriers showed higher DBP than GG genotype.

Conclusion The present study evidenced that rs1799983 NOS3 polymorphism could be associated with hypertension and DBP among Southern Europeans, being this association influenced by dietary fat (SFA and MUFA) and body mass index.

Keywords

Hypertension; Blood pressure; NOS3; Saturated fatty acids; Monounsaturated fatty acids; Obesity

314 Results

1. INTRODUCTION

Hypertension contributes to the burden of heart disease, stroke, kidney failure and premature mortality [1]. In fact, hypertension complications account for 9.4 million deaths worldwide every year [2]. Key risk factors include age, race, endocrine and metabolic disorders, lifestyle behaviors and genetics, among others [1].

In this context, familial and twin studies have estimated the heritable component of blood pressure (BP) to be about 30%-60% [3]. Moreover, genome wide association studies (GWAS) have identified a large number of polymorphisms associated with BP or hypertension, which are located in or near genes involved in the renin-angiotensin-aldosterone system, related to enzymes and receptors of the mineral-and glucocorticoid pathways and associated with proteins implicated in the structure and/or regulation of vascular tone [4].

Among them, the nitric oxide synthase (NOS3) gene is regarded as one of the putative candidate genes for BP regulation and hypertension, since it is involved in the production of nitric oxide (NO) which has vasodilator effects (i.e. inhibiting vascular smooth muscle contraction) [5]. Indeed, it has been observed in an animal model that the disruption of NOS3 gene led to hypertension, while in humans the inhibition of NOS3 elevated BP [5, 6]. Between NOS3 genetic variants, the rs1799983 is the most recognized polymorphism related not only to BP and hypertension, but also to coronary artery and vascular diseases, myocardial infarction, metabolic syndrome and type 2 diabetes [7]. Unfortunately, the results of studies seeking associations of rs1799983 and BP or hypertension have not always been consistent in different populations, which might be due to gene-environmental risk factors interactions [8]. However, to our knowledge, there are few reports on the modulation of environmental factors such as excess body weight or diet, two risk factors widely associated with increased BP levels and hypertension, on the association between rs1799983 and BP or hypertension [9, 10].

Therefore, the aims of the present research were to examine the potential association between the rs1799983 NOS3 genetic variant and BP levels and hypertension, and to investigate the possible influence of non-genetic risk factors on that association.

2. METHODS

2.1. Study population

The present study encompassed men and women of Caucasian ancestry who voluntarily attended community pharmacies located in 7 regions of Spain (Barcelona, Zaragoza, La

315 Chapter 11

Coruña, Pontevedra, Madrid, Granada and Málaga). Genotype information of 718 individuals was available. Of these, 13 subjects were excluded with missing values for dietary intake, physical activity, anthropometric measurements and/or BP levels. Therefore, the screened group included 705 individuals with a mean age of 50.2±13.2 y.o. Of the total population 22.3% (n 157) were male and 77.7% (n 548) were female.

The recruited subjects were specifically asked whether they would be willing to take part anonymously in the research study. Only those who gave written informed consent for participation were enrolled. All procedures were in accordance with the guidelines laid down in the Declaration of Helsinki. Patient data were codified to guarantee anonymity. The Research Ethics Committee of the University of Navarra provided confirmation of fulfillment of the ethical standards affecting this research (Ref. 2410/2014).

Anthropometric measurements, habitual dietary intake and physical activity were collected by trained nutritionists as described elsewhere [11]. Briefly, weight, height, waist circumference and body fat mass were measured with a digital scale (Tanita BF-522W, Tanita Corporation, Tokyo, Japan), a portable stadiometer (Leicester Tanita), inextensible tape measure and bioelectrical impedance (Tanita BF-522W), respectively. Dietary intake was determined using a food frequency questionnaire in which basic foods were classified into 19 food groups, where four responses were possible (daily, weekly, monthly or never). Physical activity was collected by a short 24-h physical activity questionnaire [12].

2.2. Blood pressure

BP was measured using the following standardized protocol with a validated automatic device (MIT Elite Plus, OMRON Healthcare, Hoofddorp, Netherland) and appropriately sized cuff [13]. Measurements were carried out in the non-dominant arm, with the elbow at the level of the right atrium and with the subject in a sitting position. Systolic BP (SBP) and diastolic BP (DBP) were taken two times, separate of at least 10 minutes. The last measurement was used in the analysis, discarding the first one. If in the second reading the SBP or DBP were ≥140 mmHg or ≥90 mmHg, respectively, was performed a third measurement.

BP was considered as a categorized variable (hypertension vs normotension) and as a continuous variable. In order to dichotomize variables, individuals with a SBP ≥140 mmHg or DBP ≥90 mmHg or with declared diagnostic hypertension were defined as “hypertensive”. On the other hand, as a continuous variable, BP was treated after adjustment of the BP treatment. In this sense, according to previous studies and based on the known average treatment

316 Results effects, fixed increment of 15 mmHg SBP and 10 DBP was added to the pressures of individuals with diagnostic hypertension [14-17].

2.3. DNA isolation and genotyping

Genomic DNA was isolated from oral epithelial cells (collected by ORAcollect DNA®, DNAGenotek, Kanata, Canada) by QIAcube using QiAmp DNA Mini QIAcube Kit (Qiagen, Hilden, Germany), following the manufacturer’s standard protocol. rs1799983 NOS3 genetic variant was genotyped by Luminex® 100/200TM System (Luminex Corporation, Austin, Texas), which is based on the principles of xMAP® Technology. This method uncompressed polystyrene microspheres internally dyed with various ratios of spectrally distinct fluorochromes, that are detected by a flow cytometry-based instrument [18].

2.4. Statistical analysis

Examination of Hardy-Weinberg equilibrium (HWE) was assessed using the chi-square test, and allele frequencies were estimated. Quantitative variables are expressed as means (SD), and differences among genotype groups were analyzed using general linear models of analysis of covariance (ANCOVA) with age, gender, energy intake and physical activity as covariates. Multiple logistic regression (categorical variable) and multiple linear regression models (continuous variables), adjusted for covariates (model A adjusted for non-modifiable factors and model B adjusted for model A plus modifiable factors), were applied to investigate the association of the NOS3 polymorphism with hypertension and SBP/DBP, respectively. Interactions between rs1799983 (as a dummy variable) and weight status or dietary fatty acid intake on BP were investigated using likelihood ratio test, once adjusted for potential confounders including gender, age, physical activity, energy intake and alcohol consumption. All statistical procedures were performed using STATA/SE version 12.0 (StataCorp, College Station, Tex., USA). Tests were considered statistically significant at p value <0.05.

3. RESULTS

Minor allele frequency (MAF) (T=36.9) and genotypic frequencies (GG=40.1, GT=46.0, TT=13.9) for rs1799983 were in accordance with the reference data for Caucasian populations (HapMap CEU). Moreover, the distribution of the polymorphism was in Hardy-Weinberg equilibrium (HWE) (p>0.05). Baseline characteristics of the population have been described according to

317 Chapter 11 rs1799983 NOS3 genotype subgroups (Table 1). There were no statistically significant differences based on genotype for anthropometrical, physical activity and dietary variables. Of the 705 individuals, 22.3% (n 157) self-declared hypertension. Moreover, about 30.6 % of the individuals self-declared that they suffered one or more metabolic disorders: 3.3 % type two diabetes, 28.6 % different lipid metabolism impairments and 3.1 % cardiovascular disease.

Logistic regression models to estimate odds ratios of self-declared hypertension for individuals with the rs1799983 polymorphism were analyzed (Fig. 1). Once adjusted for covariates, those subjects with the TT genotype showed a statistically significant 2.30-fold higher predisposition of hypertension (95%CI 1.18-4.44; p 0.014) than those subjects with the GG genotype. However, individuals with the GT genotype did not show a higher risk of hypertension than those with the GG genotype (95%CI 0.82-2.09; p 0.266). Additive and recessive model showed a statistically significant 1.46 (95%CI 1.06-2.01; p 0.019) fold per risk allele and 1.98 (95%CI 1.08-3.63; p 0.028) fold higher predisposition of hypertension, respectively. There were no a higher risk of hypertension when the dominant model was applied (OR 1.48 95% CI 0.95-2.30; p 0.084). When the hypertension was defined according to blood pressure measurements of the study, once adjusted for hypertensive treatment, no statistically significant differences were observed.

Results from linear regression analyses showed an association between rs1799983 and DBP. Nevertheless, no relationship with SBP was found (Table 2). Carriers of the T allele had significantly greater values of DBP and DBP corrected according to treated hypertensive subjects after adjusted for non-modifiable risk factors of hypertension and both non- modifiable and modifiable risk factors of hypertension. Specifically each additional risk allele was associated with a 1.99 mmHg increase in DBP, when DBP was adjusted according to treated hypertensive subjects and after statistical analysis was adjusted for gender, age, smoking, alcohol consumption and physical activity. Moreover, GT and TT individuals showed a 2.55 mmHg higher DBP than GG individuals after adjusted for non-modifiable and modifiable covariates. Upon stratifying data by gender, similar results were found among men (data not shown).

After data were adjusted for potential confounder variables, saturated fatty acids (SFA) and monounsaturated fatty acids (MUFA) modified the effect of the rs1799983 NOS3 genetic variant on DBP (Fig. 2). The influence of Body Mass Index (BMI) on the association between rs1799983 and DBP was also analyzed (Fig. 3). When the analysis was carried out according to the three genotype groups, no significant interactions were found. In contrast, when GT and TT genotype were clustered, a significant interaction was found. Among normal weight

318 Results individuals no significant differences were found in DBP between GG genotype subjects and GT and TT genotype subjects. However, among overweight individuals GT and TT genotype subjects showed higher DBP than TT genotype subjects.

4. DISCUSSION

The current study provides evidence that rs1799983 NOS3 genetic variant is associated with an increased odds of hypertension and a high to normal DBP in Southern Europeans. Interestingly, an interaction between BMI and dietary fat intake (SFA and MUFA), with the rs1799983 polymorphism to influence DBP levels was observed.

In the last few years, increased attention was paid to this NOS3 polymorphism since it is directly involved in BP regulation through NO levels [5]. A meta-analysis involving 45,287 subjects identified that the T allele of the rs1799983 polymorphism was associated with hypertension predisposition with an odds ratio of 1.20 (p 0.015) [8]. However, when the population was stratified for ethnicity, no significance was reached for Whites (OR 0.99, p 0.828). It should be highlighted that most of the studies in European population, included in the meta-analysis, were carried out in North and Central Europe. In line of our results, a study in American white women revealed a hazard ratio of hypertension for rs1799983 NOS3 genetic variant of 1.05 (p 0.03) [19]. Afterward, reports linked the rs1799983 not only to hypertension, but also to left ventricular hypertrophy, coronary artery disease and venous thromboembolism, among others [20-22]. In the rs1799983 polymorphism a guanine/thymine substitution at exon seven leads to a glutamate/aspartate substitution at position 298. Since this genetic variant alters the primary structure of the protein and could alter one or more functional properties of the enzyme, several mechanistic studies have been carried out [23- 27]. Different reports have revealed that in the presence of a T instead of a G at nucleotide position 894, NOS3 encodes a protein which leads to a higher susceptibility to cleavage into a 100-kDa fragment [23, 24]. Thus, the cleaved fragment could decrease NOS3 activity. However, other studies concluded that this finding might be a technical artifact [25, 26]. On the other hand, it has been confirmed that T allele carriers have less NOS3 bound to caveolin- 1, which is a protein essential for its activation and therefore to endothelial cell NO production [27].

To our knowledge, two previous studies have suggested an influence of rs1799983 NOS3 genetic variant on DBP [28, 29]. The research by Kimura et al., in African-derived Brazilian population, revealed a single two locus effect between rs1799983 and rs1801058, NOS3 and

319 Chapter 11

GKR4 variants, respectively, on DBP [28]. Lately, Nunes et al. showed that GG genotype subjects presented lower exercise DBP than T allele carriers, among Brazilian women [29]. Nevertheless, Seidlerová et al. [30] failed to confirm in Europeans the relationship between rs1799983 and any arterial properties. It should be mentioned that Delgado-Lista et al. [31] reported that after a meal rich in SFA the microvascular endothelial function, which is a subclinical condition found in most patients with hypertension, was lower among TG and GG genotype subjects than TT genotype subjects of a Spanish population. In one hand, this finding suggests a possible ethnic influence on genetic modulation of BP levels responses. On the other hand, as other authors have previously suggested gene-environmental interactions should be analyzed [8].

In addition to high sodium intake the scientific literature has established a link between dietary fat and increased risk of hypertension [9]. In this sense, we found gene-diet interactions between rs1799983 NOS3 genetic variant and SFA and MUFA intakes and the effect on DBP. This outcome diverges from the results of Kingah et al. that did not find any interaction between the NOS3 genetic variant and dietary fat (total fat, SFA, MUFA and polyunsaturated fatty acids (PUFA)) [32]. Nevertheless, our observations are consistent with the results of Pereira et al. [33] that suggested a possible gene-diet interaction between rs1799983 and a diet rich in SFA and cholesterol to influence BP levels. However, the mechanisms by which dietary fat modulates the effect of NOS3 genetic variant on BP levels are unknown and can at best only be speculated. In line of our results, it has been reported in a molecular model that a diet rich in olive oil (high MUFA diet) increased the expression of the NOS3 enzyme. In contrast, a diet rich in SFA decreased the expression of NOS3 [34]. Thus, we hypothesized that carriers of the T allele presented lower effect to dietary fat due to a reduced function of NOS3. To establish the mechanism by which dietary fat intake influences BP depending on the NOS3 genotype, more studies are needed.

Interestingly, this study has found a novel interaction between BMI and rs1799983 on the modulation of DBP levels. This observation is consistent with the results of Abdel-Aziz et al. [11], who described that the association of obesity with TT genotype of rs1799983 polymorphism increased the risk to develop premature coronary artery disease in Egyptians. Moreover, previous studies have identified the influence of obesity on the relationship between different polymorphisms and BP levels or hypertension among adults and children [35-37]. Initial epidemiological studies suggesting a relationship between obesity and BP levels have been supported with the understanding of potential mechanisms involved in both conditions, such as vascular and systemic insulin resistance, dysfunction of the sympathetic

320 Results nervous system and the renin-angiotensin-aldosterone system, among other pathogenetic factors [10, 38]. On the other hand, NOS3 protein content was shown to be significantly lower in overweight and obese people and inversely associated with body fat mass [39, 40]. Moreover, oxidative stress, a well-known process implicated in obesity, has been related to NOS3 response. In this regard, a previous study in an animal model suggested that a reduced NO production and an increased production of reactive oxygen species (ROS) may contribute to hypertension in obese rats [41]. Therefore, such scientific evidence suggests that our result may be a genuine interaction rather than a random chance finding.

The present study bears some strengths and limitations that need to be mentioned. First, this study included only subjects of Caucasian ancestry, so our findings may not be generalizable to other ethnic populations. Second, the study site could be considered as a random effect. However, when the statistical analysis was carried out adjusted for the covariate region the analysis gave similar results (data not shown). Third, due to the relatively small sample size and the use of a nominal significance threshold p value <0.05, the findings may need to be replicated in other population. Nevertheless, it was large enough to provide us adequate statistical power and the associations and interactions have reached statistical significance in this context of limited sample size. Fourth, we did not exclude subjects with diagnosed hypertension. However, the adjustments of SBP and DBP according to diagnostic hypertension avoided the possibility that a biased selection might result from selecting only individuals without hypertension. This correction has been widely adopted for different authors [14-17]. Fifth, non-genetic factors could influence BP levels. In this sense, multivariate adjustments support that the effects are independent of some common traditional hypertension risk factors, including age, BMI, physical activity, smoking and alcohol consumption. However, sodium intake was not taken into account, although it would be of interest to be included in the analysis, but unfortunately data were not available. Finally, as a cross-sectional study, to get definitive clinical conclusions about dietary intake may be elusive.

The present research represents a pilot effort to explore the role of gene-environment interactions in BP levels. The investigation of genetic factors and their interactions with the environment may improve the selection of more individualized effective treatment of hypertension.

321 Chapter 11

Acknowledgements

We thank all the seekers of the Nutrigenetic Service who voluntarily offered their data. The authors are grateful to the 5 nutritionists for data collection and to Amaia Ibañez for excellent technical assistance. The authors thank the Linea Especial (University of Navarra; LE/97) and to the Spanish Ministry of Economy and Competitiveness (AGL2013-45554-R project) for financial support, CIBER Fisiopatología de la Obesidad y Nutrición (CIBERobn)/RETICS schedules (Instituto de Salud Carlos III, Madrid, Spain) for assistance in this study and the support from CINFA (Olloki, Spain) concerning the genetic tools and general logistic. The pre-doctoral research grant to Leticia Goñi from the Asociación de Amigos Universidad de Navarra is gratefully acknowledged.

Ethical standards

The Research Ethics Committee of the University of Navarra provided confirmation of fulfillment of the ethical standards affecting this research (ref. 2410/2014). Therefore, the survey was in accordance with the principles of the 1964 Declaration of Helsinki and its later amendments.

Conflict of interest

The authors declare that they have no conflict of interest

322 Results

REFERENCES

1. World Health Organization (WHO) (2013) A global brief on hypertension. World Health Organization, Geneva 2. Lim SS, Vos T, Flaxman AD, et al (2012) A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 380:2224-2260 3. Zhao Q, Kelly TN, Li C, He J (2013) Progress and future aspects in genetics of human hypertension. Curr Hypertens Rep 15:676-686 4. Butler MG (2010) Genetics of hypertension. Current status. J Med Liban 58:175-178 5. Gamboa A, Shibao C, Diedrich A, Choi L, Pohar B, Jordan J, Paranjape S, Farley G, Biaggioni I (2007) Contribution of endothelial nitric oxide to blood pressure in humans. Hypertension 49:170-177 6. Huang PL, Huang Z, Mashimo H, Bloch KD, Moskowitz MA, Bevan JA, Fishman MC (1995) Hypertension in mice lacking the gene for endothelial nitric oxide synthase. Nature 377:239-242 7. Vecoli C (2014) Endothelial nitric oxide synthase gene polymorphisms in cardiovascular disease. Vitam Horm 96:387-406 8. Niu W, Qi Y (2011) An updated meta-analysis of endothelial nitric oxide synthase gene: three well-characterized polymorphisms with hypertension. PLoS One 6:e24266 9. Hall WL (2009) Dietary saturated and unsaturated fats as determinants of blood pressure and vascular function. Nutr Res Rev 22:18-38 10. DeMarco VG, Aroor AR, Sowers JR (2014) The pathophysiology of hypertension in patients with obesity. Nat Rev Endocrinol 10:364-376 11. Abdel-Aziz TA, Mohamed RH (2013) Association of endothelial nitric oxide synthase gene polymorphisms with classical risk factors in development of premature coronary artery disease. Mol Biol Rep 40:3065-3071 12. Panel on Macronutrients, Panel on the Definition of Dietary Fiber, Subcommittee on Upper Reference Levels of Nutrients, Subcommittee on Interpretation and Uses of Dietary Reference Intakes, and the Standing Committee on the Scientific Evaluation of Dietary Reference Intakes (2005) Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids (Macronutrients). The National Academies Press, Washington, D.C. 13. Grim CE, Grim CM (2009) The Omron Elite 7300W home blood pressure monitor passes the European Society of Hypertension International Validation Protocol for women and men. Blood Press Monit 14:87-90 14. Fava C, Sjogren M, Montagnana M, Danese E, Almgren P, Engstrom G, Nilsson P, Hedblad B, Guidi GC, Minuz P, Melander O (2013) Prediction of blood pressure changes over time and incidence of hypertension by a genetic risk score in Swedes. Hypertension 61:319-326 15. International Consortium for Blood Pressure Genome-Wide Association Studies, Ehret GB, Munroe PB, et al (2011) Genetic variants in novel pathways influence blood pressure and cardiovascular disease risk. Nature 478:103- 109 16. Levy D, Ehret GB, Rice K, Verwoert GC, Launer LJ, Dehghan A, Glazer NL, Morrison AC, Johnson AD, Aspelund T, Aulchenko Y, Lumley T, Kottgen A, Vasan RS, Rivadeneira F, Eiriksdottir G, Guo X, Arking DE, Mitchell GF, Mattace- Raso FU, Smith AV, Taylor K, Scharpf RB, Hwang SJ, Sijbrands EJ, Bis J, Harris TB, Ganesh SK, O'Donnell CJ, Hofman A, Rotter JI, Coresh J, Benjamin EJ, Uitterlinden AG, Heiss G, Fox CS, Witteman JC, Boerwinkle E, Wang TJ, Gudnason V, Larson MG, Chakravarti A, Psaty BM, van Duijn CM (2009) Genome-wide association study of blood pressure and hypertension. Nat Genet 41:677-687

323 Chapter 11

17. Newton-Cheh C, Johnson T, Gateva V, et al (2009) Genome-wide association study identifies eight loci associated with blood pressure. Nat Genet 41:666-676 18. Dunbar SA (2006) Applications of Luminex xMAP technology for rapid, high-throughput multiplexed nucleic acid detection. Clin Chim Acta 363:71-82 19. Conen D, Cheng S, Steiner LL, Buring JE, Ridker PM, Zee RY (2009) Association of 77 polymorphisms in 52 candidate genes with blood pressure progression and incident hypertension: the Women's Genome Health Study. J Hypertens 27:476-483 20. Rai H, Parveen F, Kumar S, Kapoor A, Sinha N (2014) Association of endothelial nitric oxide synthase gene polymorphisms with coronary artery disease: an updated meta-analysis and systematic review. PLoS One 9:e113363 21. Xin Y, Song X, Xue H, Liu Z, Wang X, Wang H, Sun K, Bai Y, Liu J, Hui R (2009) A common variant of the eNOS gene (E298D) is an independent risk factor for left ventricular hypertrophy in human essential hypertension. Clin Sci (Lond) 117:67-73 22. Qin J, Dai J, Xu Z, Chen D, Qin J, Shi D, Teng H, Jiang Q (2013) Genetic polymorphism of NOS3 with susceptibility to deep vein thrombosis after orthopedic surgery: a case-control study in Chinese Han population. PLoS One 8:e70033 23. Persu A, Stoenoiu MS, Messiaen T, Davila S, Robino C, El-Khattabi O, Mourad M, Horie S, Feron O, Balligand JL, Wattiez R, Pirson Y, Chauveau D, Lens XM, Devuyst O (2002) Modifier effect of ENOS in autosomal dominant polycystic kidney disease. Hum Mol Genet 11:229-241 24. Tesauro M, Thompson WC, Rogliani P, Qi L, Chaudhary PP, Moss J (2000) Intracellular processing of endothelial nitric oxide synthase isoforms associated with differences in severity of cardiopulmonary diseases: cleavage of proteins with aspartate vs. glutamate at position 298. Proc Natl Acad Sci U S A 97:2832-2835 25. Fairchild TA, Fulton D, Fontana JT, Gratton JP, McCabe TJ, Sessa WC (2001) Acidic hydrolysis as a mechanism for the cleavage of the Glu(298)-->Asp variant of human endothelial nitric-oxide synthase. J Biol Chem 276:26674- 26679 26. McDonald DM, Alp NJ, Channon KM (2004) Functional comparison of the endothelial nitric oxide synthase Glu298Asp polymorphic variants in human endothelial cells. Pharmacogenetics 14:831-839 27. Joshi MS, Mineo C, Shaul PW, Bauer JA (2007) Biochemical consequences of the NOS3 Glu298Asp variation in human endothelium: altered caveolar localization and impaired response to shear. FASEB J 21:2655-2663 28. Kimura L, Angeli CB, Auricchio MT, Fernandes GR, Pereira AC, Vicente JP, Pereira TV, Mingroni-Netto RC (2012) Multilocus family-based association analysis of seven candidate polymorphisms with essential hypertension in an african-derived semi-isolated brazilian population. Int J Hypertens 2012:859219 29. Nunes RA, Barroso LP, Pereira Ada C, Krieger JE, Mansur AJ (2014) Gender-related associations of genetic polymorphisms of alpha-adrenergic receptors, endothelial nitric oxide synthase and bradykinin B2 receptor with treadmill exercise test responses. Open Heart 1:e000132-2014-000132. eCollection 2014 30. Seidlerova J, Filipovsky J, Mayer O,Jr, Kucerova A, Pesta M (2015) Association between endothelial NO synthase polymorphisms and arterial properties in the general population. Nitric Oxide 44:47-51 31. Delgado-Lista J, Garcia-Rios A, Perez-Martinez P, Fuentes F, Jimenez-Gomez Y, Gomez-Luna MJ, Parnell LD, Marin C, Lai CQ, Perez-Jimenez F, Ordovas JM, Lopez-Miranda J (2011) Gene variations of nitric oxide synthase regulate the effects of a saturated fat rich meal on endothelial function. Clin Nutr 30:234-238 32. Kingah PL, Luu HN, Volcik KA, Morrison AC, Nettleton JA, Boerwinkle E (2010) Association of NOS3 Glu298Asp SNP with hypertension and possible effect modification of dietary fat intake in the ARIC study. Hypertens Res 33:165-169

324 Results

33. Pereira TV, Rudnicki M, Cheung BM, Baum L, Yamada Y, Oliveira PS, Pereira AC, Krieger JE (2007) Three endothelial nitric oxide (NOS3) gene polymorphisms in hypertensive and normotensive individuals: meta-analysis of 53 studies reveals evidence of publication bias. J Hypertens 25:1763-1774 34. Martins MA, Catta-Preta M, Mandarim-de-Lacerda CA, Aguila MB, Brunini TC, Mendes-Ribeiro AC (2010) High fat diets modulate nitric oxide biosynthesis and antioxidant defence in red blood cells from C57BL/6 mice. Arch Biochem Biophys 499:56-61 35. Grove ML, Morrison A, Folsom AR, Boerwinkle E, Hoelscher DM, Bray MS (2007) Gene-environment interaction and the GNB3 gene in the Atherosclerosis Risk in Communities study. Int J Obes (Lond) 31:919-926 36. Xi B, Zhao X, Chandak GR, Shen Y, Cheng H, Hou D, Wang X, Mi J (2013) Influence of obesity on association between genetic variants identified by genome-wide association studies and hypertension risk in Chinese children. Am J Hypertens 26:990-996 37. Marteau JB, Sass C, Pfister M, Lambert D, Noyer-Weidner M, Visvikis S (2004) The Leu554Phe polymorphism in the E-selectin gene is associated with blood pressure in overweight people. J Hypertens 22:305-311 38. Silva DA, Petroski EL, Peres MA (2013) Accuracy and measures of association of anthropometric indexes of obesity to identify the presence of hypertension in adults: a population-based study in Southern Brazil. Eur J Nutr 52:237-246 39. Hickner RC, Kemeny G, Stallings HW, Manning SM, McIver KL (2006) Relationship between body composition and skeletal muscle eNOS. Int J Obes (Lond) 30:308-312 40. Kraus RM, Houmard JA, Kraus WE, Tanner CJ, Pierce JR, Choi MD, Hickner RC (2012) Obesity, insulin resistance, and skeletal muscle nitric oxide synthase. J Appl Physiol (1985) 113:758-765 41. da Cunha NV, Pinge-Filho P, Panis C, Silva BR, Pernomian L, Grando MD, Cecchini R, Bendhack LM, Martins-Pinge MC (2014) Decreased endothelial nitric oxide, systemic oxidative stress, and increased sympathetic modulation contribute to hypertension in obese rats. Am J Physiol Heart Circ Physiol 306:H1472-80

325 Chapter 11

Table 1 Characteristics of the population by NOS3 rs1799983 genotype

Total GG GT TT p-value mean (SD) mean (SD) mean (SD) mean (SD) n (%) 705 283 (40.1) 324 (46.0) 98 (13.9) 0.733a Women (%) 548 (77.7) 211 (29.9) 260 (36.9) 77 (10.9) - Age (years)b 50.2 (13.2) 50.6 (13.1) 50.7 (13.1) 47.7 (13.1) 0.121 Anthropometric measurements Height (cm) b 163.0 (8.7) 163.0 (6.6) 163.0 (6.6) 163.2 (6.6) 0.886 Weight (kg)c 78.3 (16.7) 77.9 (13.6) 79.3 (13.6) 75.9 (13.7) 0.083 BMI c 29.5 (5.8) 29.3 (5.0) 29.8 (5.0) 28.7 (5.1) 0.155 BFM (%)c 34.6 (10.2) 34.6 (7.4) 34.7 (7.4) 34.2 (7.4) 0.861 Waist circumference (cm) c 96.5 (15.2) 95.8 (12.1) 97.4 (12.1) 95.2 (12.1) 0.157 Physical activity c Physical activity level c 1.23 (0.03) 1.23 (0.03) 1.23 (0.03) 1.24 (0.03) 0.153 Baseline dietary intake Energy (kcal/day) d 2,149 (430) 2,138 (391) 2,150 (392) 2,144 (393) 0.791 Carbohydrate (g/day) e 197.8 (71.4) 197.5 (53.3) 197.8 (53.3) 198.6 (53.4) 0.986 Protein (g/day) e 92.9 (24.8) 93.2 (17.0) 92.7 (17.0) 92.8 (17.0) 0.962 Fat (g/day) e 95.1 (22.4) 95.4 (16.6) 95.2 (16.7) 94.4 (16.7) 0.895 MUFA (g/day) e 46.6 (11.8) 46.9 (9.9) 46.5 (9.9) 45.9 (9.9) 0.641 PUFA (g/day) e 14.1 (3.7) 14.1 (2.6) 14.1 (2.6) 14.0 (2.6) 0.889 SFA (g/day) e 19.1 (5.4) 19.0 (4.0) 19.3 (4.0) 19.1 (4.0) 0.681 BMI, Body mass index; BFM, Body fat mass; %E, Percentage of energy; MUFA; Monounsaturated fatty acids; Polyunsaturated fatty acids; SFA, Saturated fatty acids a Hardy Weinberg Equilibrium b Adjusted for gender c Adjusted for gender, age and energy intake d Adjusted for gender, age and physical activity e Adjusted for gender, age, energy intake and physical activity

326 Results

Table 2 Association of the rs1799983 NOS3 genetic variant with Systolic and Diastolic BP among Southern Europeansa

Model 1 Model 2 Genotype B (95% CI) p B (95% CI) p SBP Additive effect of T allele 1.06 (-0.95-3.07) 0.303 1.25 (-0.69-3.19) 0.205 GT vs GG 0.59 (-2.39-3.58) 0.696 -0.00 (-2.87-2.86) 0.998 TT vs GG 2.43 (-1.87-6.73) 0.267 3.38 (-0.76-7.53) 0.110 GT + TT vs GG 1.02 (-1.80-3.84) 0.478 0.78 (-1.93-3.49) 0.573 TT vs GG + GT 2.11 (-1.88-6.11) 0.299 3.38 (-0.47-7.24) 0.085 DBP Additive effect of T allele 1.81 (0.31-3.32) 0.018 1.99 (0.61-3.36) 0.005 GT vs GG 2.87 (0.63-5.10) 0.012 2.16 (0.13-4.20) 0.037 TT vs GG 2.90 (-0.32-6.12) 0.077 3.86 (0.91-6.80) 0.010 GT + TT vs GG 2.88 (0.77-5.00) 0.008 2.55 (0.63-4.47) 0.009 TT vs GG + GT 1.37 (-1.63-4.38) 0.370 2.71 (-0.04-5.45) 0.053 SBP, Systolic blood pressure; DBP, Diastolic blood pressure; 95% CI, 95% confidence interval a SBP + 15mmHg and DBP + 10mmHg to treated hypertensive subjects Model 1 adjusted for gender and age Model 2 adjusted for gender, age, BMI, smoke, physical activity and alcohol consumption

327 Chapter 11

Figure 1 Risk of hypertension according to rs1799983 genotype Adjusted for gender, age, physical activity, energy intake, smoke and alcohol consumption

328 Results

Figure 2 Interaction between NOS3 and SFA (a) and MUFA (b) on DBP DBP, Diastolic blood pressure; SFA, Saturated fatty acids; MUFA, Monounsaturated fatty acids Adjusted for gender, age, physical activity, energy intake, smoke and alcohol consumption

329 Chapter 11

Figure 3 Interaction between NOS3 and obesity status on DBP Adjusted for gender, age, physical activity, energy intake, smoke and alcohol consumption

330

GENERAL DISCUSSION

General discussion

1. JUSTIFICATION OF THE STUDY

The rates of non-communicable diseases have reached epidemic proportions worldwide, with obesity and metabolic related disorders being among the most prevalent (271). For this reason, the prevention and treatment of these diseases, as well as the investigation of the potential associations between them, has attracted much attention in recent years.

As a result, a number of studies have analyzed different weight loss strategies aimed at reducing energy intake and increasing energy expenditure to combat the obesity epidemic and accompanying consequences (193). However, the multifactorial nature of obesity has made it difficult to develop effective therapies that guarantee not only the body weight loss but also the body weight lowering maintenance. Because of the interindividual variability in the response to weight loss interventions, recent research has raised interest in other non- traditional factors (dietary habits, physical activity, bioactive components), such as genetic, epigenetic and metagenomic data as we ascertained in our previous review (197).

Gene-environment interactions may also take part in the management of obesity and metabolic syndrome features (419). In this sense, there is information concerning the potential effect of the diet on the association between a genetic variant and a phenotype (168). Therefore, the study of the influence of genetic variations on the body’s response to nutrients, are assuming a new model of obesity treatment and associated comorbidities through personalized diets, according to the genetic predisposition of the individual. In this sense, it is important to assess adequately the dietary habits of each individual. For this reason, in the present work we studied the validity of a FGFQ, based on 19 groups of foods, to demonstrate that it measure what is desired effectively. Although the FGFQ of the present study could be classified as a short questionnaire, because it is based on a food exchange system indirectly reflects the assessment of a greater number of foods (390).

Consequently, the present study was devised to assess possible genetic factors underlying obesity, weight loss and metabolic syndrome features in different nutritional intervention studies, evaluating also the impact of dietary factors in the association between genetic and phenotypic features.

2. GENETIC SUSCEPTIBILITY TO OBESITY

First, twin and adoption studies demonstrate that the tendency to be obese is highly heritable (31). By 2004, candidate gene studies and GWAS have identified over 600 candidate genes or

333 General discussion chromosomal regions implicated in the pathogenesis of obesity (420). Despite this, much of the heritability in obesity remained to be explained.

The rs9939609 FTO variant is the most well-known polymorphism associated with obesity. The FTO gene has a curious early history of association with obesity. In 2007, through a GWAS FTO was related to type 2 diabetes. However, when the analysis was adjusted for BMI, the association between FTO polymorphism and type 2 diabetes disappeared. This finding suggested that FTO polymorphisms were primarily affecting BMI rather type 2 diabetes. After that, FTO became the first gene to emerge from GWAS unequivocally linked to obesity (64– 66). As other authors have found, in the present work it was observed that individuals with the FTO rs9939609 AA genotype had greater BMI, percentage of body fat mass, waist circumference and waist-to-height ratio. Moreover, after the LARS analysis FTO rs9939609 was selected as a polymorphism involved in BMI explanation. FTO is a nuclear protein member of the AlkB related non-haem iron and 2-oxoglutarate-dependent oxygenase superfamily (421). Although the relationship between FTO genetic variant and obesity related traits (BMI, obesity risk, waist circumference, body fat mass) has been confirmed in several populations, the physiological function of this gene in body weight regulation seems unclear (422). Results from animal and human studies suggest that FTO gene acts on food intake with no impact on resting energy expenditure, although in rodents results are controversial (423). Since Fto in rodents is widely expressed in the brain, including the hypothalamic nucleic, it has been proposed that FTO could be linked to BDNF-NTRK2 signaling pathway influencing food intake regulation (424). Moreover, Fto has been implicated in the regulation of leptin levels and leptin sensitivity (425,426). Other mechanism of action could be as a sensor of amino acids availability (427). Additionally, SNPs in FTO could exerts their effects in other genes including RLGRIPIL and IRX3 (428,429).

After the discovery of FTO, other genetic variants have been related to adiposity traits (40,420). Among them, PPARG has been associated to obesity in different candidate gene studies and GWAS (50,128). PPARG is member of the nuclear hormone receptor subfamily of ligand-dependent transcription factors and it is primarily expressed in adipocytes where it modulates the expression of target genes involved in adipocyte differentiation, insulin sensitivity and inflammatory processes (430,431). Of the several variants identified in the PPARG gene, the most studied variant in relation to adiposity and insulin resistance is Pro12Ala (rs1801282), that has been associated with reduced ability to transactivate responsive promoters and, thus, with lower transcriptional activity (50,432,433). The result of the LARS analysis suggested that PPARG rs1801282 was negatively associated with BMI. These results

334 General discussion are against of most candidate gene studies in which Pro12Ala has been associated with higher BMI; however, other authors reported the opposite association or have not found any association at all (50,56,433–436). These controversial results suggest that, if this variant does influence obesity predisposition, it may do so through environment-dependent mechanisms. In fact, several studies have reported interactions between PPARG and environmental factors such as gender, dietary fat intake or breast feeding on obesity traits (128,437–439).

Although the association between genetic variants of the PPARG gene and obesity traits has been widely studied, as far as we know there is limited evidence regarding the relationship between PPARA variants, another gene that encodes for a nuclear transcription factor, and obesity phenotype. PPARA is expressed in cells with high rates of fatty acid catabolism including liver, heart, kidney and skeletal muscle to regulate fatty acid oxidation systems (431). In our research although, there was no a significant association between PPARA rs1800206 and BMI, the results of the LARS analysis suggested a negatively association. In contrast to our results Costa-Urrutia et al. (2017) reported a positive association between the PPARA rs1800206 polymorphism and obesity risk and Sirbelnagel et al. (2009) did not find a relationship between such genetic variant and BMI or body fat composition (440,441). We hypothesized that our opposite results regarding PPARG and PPARA could be due partly to the fact that we have carried out the analysis in the presence of other non-genetic and genetic variants.

As far as we know our group reported for the first time a significant association between APOE rs429358 genetic variant and BMI. C allele carriers showed significantly higher BMI and a trend toward significance for body fat mass, waist circumference and hip circumference. After that, APOE rs429358 variant was included in the multivariable model regression as a significant BMI predictor. The APOE gene encodes for a structural apolipoprotein that plays a major role in maintaining plasma lipids homeostasis (442). APOE genetic variants have been associated with several metabolic disorders including high obesity risk (443–446). The contribution of APOE to obesity could be explained because of its effect on adipogenesis as has been reported in vivo and in vitro studies (447,448).

Genetic variants identified by candidate gene studies and GWAS explain surprisingly a small proportion of the phenotypic variance. For this reason, approaches that combine effects of more than one loci, rather than considering only one locus, in a GRS has attempted (145). In this context, we computed a GRS of 16 polymorphisms previously related to obesity and lipid metabolism traits. The GRS was positively associated with several measurements of obesity and obesity risk in our study population. However, the analyses demonstrated that the GRS

335 General discussion explained 0.38% of the BMI variability and 1.46% of the percentage fat mass heritability. In addition, when we generate a multivariable regression model based on 4 of the 16 polymorphisms selected by LARS they accounted for 0.5% of the BMI variability in a large population. Our findings are in accordance with the studies by Martínez-García et al. (2013), Belsky et al. (2013) and Li et al. (2010), where a small number of SNPs explained less than 1% of the BMI heritability (145,146,159). In this sense, it should be highlighted that when Locke et al. (2015) included a total of 97 SNPs in a prediction model of BMI the authors found a BMI explanation of 2.7% (72).

Because it is clear that individual risk of obesity reflects the integration of genetic and non- genetic factors, a multivariable regression model based on phenotype and genotype variables was generated. According to the LARS analysis age, physical activity, energy intake, FTO rs9939609 variant, APOE rs429358 variant, PPARG rs1801282 variant and PPARA rs1800206 variant were involved in BMI explanation. Interestingly, 21% of the phenotypic variance in BMI was explained by the regression model obtained by LARS. Prediction models that added energy intake and physical activity have not been reported up to date, so we cannot be able to compare our results. However, some authors have observed that, when phenotypical factors are included in the genetic model (such as socioeconomic or depression status), the percentage of the explanation of the BMI significantly increases (165,449).

Several potential explanations can be offered for the low predictive value of the GRSs, but are mainly related to the fact that obesity is characterized for being a multifactorial disease (450). For example, in the present study although were included in the multivariable model the two main factors that characterized obesity, energy intake and physical activity, there are other features that have not been taken into account such as social determinants (education level, economic status), endocrine disorders (hypothyroidism) or use of certain medications (7–10). Another explanation for the relatively low predictive value of the GRSs could be related with the marginal effect sizes of the tested variants and the skewed distribution of the effect sizes. In addition, predictive models could include other sources of variation known or hypothesized to influence BMI, such as rare variants, gene-gene and gene-environment interactions, copy number variation and epigenetic and metagenomic effects (197). Finally, it should be mentioned that main studies BMI instead of body fat mass is selected as dependent variable. Although BMI is the adiposity measurement most widely used in epidemiological studies, the interpretation does not differ between gender and race, and neither distinguishes between degree of fatness, muscle mass and skeletal mass (451). Therefore, it can lead to errors in the

336 General discussion estimation of adiposity, over or underestimating adiposity depending on subject complexion; such as athletes or metabolic obese normal weight individuals.

Regarding the presence of gene-environment interactions the GRS interactions with dietary components were examined. After Benjamini-Hochberg correction, the intake of energy, total protein, vegetable protein, SFA, PUFA and total carbohydrates intake modified the association between the GRS and the percentage of body fat mass and obesity risk (determined by the percentage of body fat mass). Higher total protein intake was significantly associated with higher percentage of body fat mass in both genetic risk groups, but the association was significantly stronger among individuals of the high genetic risk group. Whereas, vegetable protein intake appears to have a protective effect among individuals of the low genetic risk group. Previously, Rukh et al. (2013) found that protein intake modulates the association between a 13 polymorphisms’ GRS and obesity and fat mass (152). On the other hand, although the association between SFA intake and obesity is well studied our group for the first time confirmed the effect of this nutrient and a GRS on adiposity (452). In this context, Qi et al. (2014) reported that among individuals with a higher GRS, the association between fried food consumption and BMI was stronger than that among individuals with a lower GRS (161). Our study also suggests that PUFA intake could modify the genetic association with body composition. This is in accordance with findings from a previous study which detected a significant interaction between a 12 obesity SNPs’ GRS and omega 3 PUFA and BMI (151). Another novel finding of our study was the interaction between total carbohydrates and the GRS on percentage of body fat mass. As far as we know there are no previous reports regarding this interaction, but Qi et al. (2012) found an interaction between sugar-sweetened beverages and a GRS of 32 obesity polymorphisms (155). Interestingly, Nettleton et al. (2015) reported a nominally significant interaction between a diet score and a GRS (based on 14 variants related to waist to hip ratio) on BMI adjusted waist hip ratio in 68317 adults of European ancestry (153).

3. GENETIC VARIANTS IMPLICATED IN WEIGHT LOSS

A number of strategies have been investigated in order to induce a negative energy balance and body weight loss (3). However, individual responses to body weight loss interventions vary widely and several studies have aimed to identify psychological, behavioral and personal predictors of this variability (194–196). Genetic factors have been described to be associated with adiposity and body weight control, since there are diverse genes involved in the

337 General discussion regulation of energy expenditure, appetite, thermogenesis, adipogenesis, inflammation, insulin resistance and lipid metabolism (197). Specifically, in the present work we tested the effect of ADCY3 rs10182181 and MTNR1B rs10830963 genetic variants on body weight regulation after different weight loss strategies.

Because obesity is a multifactorial disorder, the study of genes implicated in a broad spectrum of processes could be of great interest (450). This is the case of ADCY3 gene, which is widely expressed in most cell types and it is involved in several physiological and pathophysiological functions (453). The ADCY3 gene encodes for a membrane-associated enzyme that catalyzes the formation of the secondary messenger cyclic adenosine monophosphate (cAMP) from adenosine triphosphate (ATP). This messenger is involved in a large number of physiological metabolic processes including the regulation of carbohydrate and lipid metabolisms, and the development and function of the adipose tissue regulating the expression of genes involved in adipogenesis, thermogenesis and lipolysis (453,454). Different genetic variants located near or in ADCY3 gene have been associated with obesity including candidate gene studies and GWAS (70–73,455,456). For example, the largest meta-analysis of GWAS on BMI carried out-to-date found that the ADCY3 rs10182181 and rs713586 variants were associated with BMI (72). In another GWAS the rs713586 polymorphism, which is in strong linkage disequilibrium with the rs10182181 polymorphism, was identified as an expression quantitative trait loci (eQTL) since it was related to ADCY3 gene expression in different tissues (lymphocytes, omental fat and blood) (70). Interestingly, after an in silico analysis we have confirmed that the rs10182181 polymorphism is located in the promoter region of the ADCY3 gene. Particularly the rs10182181 genetic variant is located in a binding site for transcription factors including USF1, POLR2A, BHLHE40, JUNB and CREM. This observation suggests that rs10182181 polymorphism may be important for the biological function of the ADCY3 gene.

In accordance with the studies in humans, the Adcy3 knockout mice developed obesity characterized by an increase in fat mass and larger adipocytes (457). Furthermore, the authors reported that Adcy3-/- mice exhibited reduced physical activity, increased food intake, and leptin insensitivity; being speculated that these phenotypic changes could be associated with disruption of cAMP signaling in primary cilia of the hypothalamus. Recently, the same group using a floxed Adcy3 mouse strain determined that Adcy3 in the hypothalamus regulated energy expenditure (458). Apart from the hypothalamus, it has been found that Adcy3 is over- expressed in pancreatic islets of non-obese-type 2 diabetic Goto-Kakizaki rats, playing an important role in insulin secretion regulation (459). Also there is some evidence to suggest that Adcy3 may play specific physiological roles in major depression and sleep disruption,

338 General discussion which are disorders strongly associated with the obesity phenotype (460). Moreover, Adcy3 may functionally couple to Mc4r in the hypothalamus, because activation of adenylyl cyclase activity by alpha-melanocyte stimulating hormone downstream in the leptin pathway is required for the anorectic activity of leptin (461). In fact, Mc4r and Adcy3 knockout mice exhibit similar phenotypes including obesity and hyperinsulinemia (457,462).

In the OBEKIT study, macronutrient distribution significantly modified the effect of the ADCY3 rs10182181 genetic variant on changes in anthropometric and body composition measurements. The participants with the G allele of the ADCY3 rs10182181 variant showed a greater decreased in body composition measurements when consuming a low-fat diet. In this sense, it has been reported that a high-fat diet decreased the Adcy3 expression in white adipose tissue, liver and muscle (463). This haploinsufficiency confers decreased expression of genes involved in thermogenesis, fatty acid oxidation and insulin signaling in mice, while it enhanced the expression of genes related to adipogenesis in peripheral tissues (463). Moreover, mice with a gain-of-function mutation in Adcy3 presented increased Adcy3 activity and cAMP production, and consequently the mutation protects mice from high fat diet- induced metabolic disorders (464). However, the mechanisms underlying the modulation of macronutrient intake on the ADCY3 genetic variant are not fully understood and further experimental studies are needed.

On the other hand, it has long been recognized that circadian system is implicated in the regulation of energy balance, and subsequently affecting body fatness (465). In humans beings, disruption of circadian rhythms by means of shift work, social jet lag, sleep deprivation, timed feeding and consumption of a high-fat diet among others, has been related to obesity and metabolic disturbances such as type 2 diabetes and cardiovascular disease (466–468). One of the most important chronobiotics is melatonin, a hormone secreted mainly by the pineal gland and a key mediator used by the central master clock as an internal synchronizer coordinating central and peripheral tissues (469). Recent GWAS have identified common variants in the MTNR1B gene, which encodes one of the two high-affinity receptors for melatonin (470,471), associated with fasting plasma glucose and the risk of type 2 diabetes (472–474). Among them, the MTNR1B rs10830963 risk allele has been related to obesity and adiposity traits as well as to melatonin signaling (475–478).

In this context, a possible effect of the MTNR1B rs10830963 genetic variant in regulating body weight loss in the Nutrigenetic Service Ns cohort was found. Women carrying the minor G allele showed a lower weight loss and therefore a lower BMI loss than CC subjects. These results are consistent with the biological role of melatonin in energy metabolism and energy

339 General discussion balance (469). On one hand, melatonin plays roles not only in the regulation of metabolic processes, but also in the maintenance of their circadian organization (479). On the other hand, the effect of melatonin on energy balance has been consistently reported (469). Wolden-Hanson et al. (2000) demonstrated that melatonin supplementation therapy decreased body weight and intraabdominal fat independent of food intake in rats (480). In the same study, treated rats showed an increase in the nocturnal locomotor activity and in the core body temperature, confirming a putative increase in energy expenditure rather than a decrease in energy intake. In addition, evidence has indicated that melatonin also plays a key role in the regulation of adipocyte biology (lipolysis, lipogenesis), the activation of brown adipose tissue and participation in the browning process of white adipose tissue (469,481).

While the precise mechanisms underlying our results remain largely unknown, several lines of evidence have implicated the MTNR1B rs10830963 genotype in regulating melatonin signaling. Thus, Tuomi et al. (2016) ascertained that subjects carrying 1 and 2 MTNR1B rs10830963 G alleles showed a 2- and 4-fold increase in MTNR1B mRNA expression in human pancreatic islets, respectively, compared with the non-carriers (478). Nonetheless, the MTNR1B rs10830963 eQTL seems to be specific for human pancreatic islets since it has not been found in any other tissues (478). Moreover, these authors demonstrated that the administration of melatonin to nondiabetic individuals inhibits insulin secretion in all subjects, and the effect was stronger among GG than those did not carry this allele. Another recent study reported that the common genetic variant was associated with the timing of the melatonin rhythm (482). MTNR1B rs10830963 G allele carriers showed a later melatonin offset and a longer duration of elevated melatonin levels. The authors suggested that the disruption of melatonin rhythm among carriers of the risk allele may result in an increase of food intake to coincide with elevated melatonin levels in the morning leading to decreased glucose tolerance.

Interestingly, in the Nutrigenetic Service Ns cohort the effect of MTNR1B variant on weight loss was modified by genetic and non-genetic factors. On one hand, in the presence of obesity related loci (risk allele carriers of both FTO and MC4R) the MTNR1B rs10830963 was significantly associated with body weight and BMI loss in the total population. However, in the non-risk allele carriers or risk allele carriers of either FTO or MC4R group, the association disappeared. This outcome suggests that polymorphisms located in FTO and MC4R at the same time influence the relationship between MTNR1B and body weight loss which may be due to a synergistic effect. This finding may partly explain the discrepancies bout the effect of FTO, MC4R and MTNR1B and body weight loss among different studies and populations. On the other hand, baseline protein and animal protein intake significantly modified the effect of the

340 General discussion

MTNR1B variant and weight loss. In this context, there is evidence that a high-protein diet affects the rhythmic expression of circadian clock genes in mouse peripheral tissues (483). It was hypothesized that those individuals with a high-protein intake at baseline had adapted to the increased thermogenic response of a high-protein diet, and this factor could inhibit the body weight loss. However, the process by which the usual protein intake influences body weight loss depending on MTNR1B genetic variant is unknown.

Although in the POUNDS Lost trial we did not observe an association between the MTNR1B rs10830963 and changes in weight, we found significant interactions between the circadian rhythm genotype and dietary fat intake on changes in fatness, fat distribution and body composition measurements at 6 months of the dietary intervention. In response to the low-fat diet, increasing number of the G allele was associated with a greater reduction in body weight, BMI, WC and total body fat. Notably, a high-fat diet has been found to alter the circadian molecular clock, both centrally and peripherally leading to a state of chronodisruption in animals (465,484). Interestingly, several previous studies reported that the expression levels of clock genes and also the rhythmic mRNA expression were influenced by a high-fat diet consumption, in different animal tissues (485,486). Moreover, it has been demonstrated that a significant alteration of circadian rhythmicity of different hormones such as pineal melatonin, leptin, ghrelin and adiponectin among others, is related to obesity in rats fed a high-fat diet (487–489). Taken together, these data lent support to potential interplays between the MTNR1B genotype and dietary fat.

4. GENETIC VARIANTS IMPLICATED IN OBESITY RELATED TRAITS

The increased prevalence of obesity worldwide has led to a concomitant increase in the prevalence of diabetes, hypertension or dyslipidemias (258). As obesity, such metabolic comorbidities are multifactorial disorders which are in part genetically influenced. In the present study, we analyzed the effect of MTNR1B rs10830963 variant on lipid levels, PPM1K rs1440581 on glucose metabolism traits and NOS3 rs1799983 on blood pressure levels.

Since there is scientific evidence that lipid metabolism is controlled by the circadian system through rhythmic activation or repression of genes involved in lipid metabolism, either directly or indirectly by controlling other transcription factors, we aimed to investigate the effect of the MTNR1B rs10830963 variant and the level of fat intake on changes in lipid metabolism traits. In one of the largest available randomized dietary intervention trials on weight loss, apparently for the first time, report a significant interaction between the circadian rhythm

341 General discussion related MTNR1B rs10830963 genetic variant and dietary fat intake on changes in total cholesterol and LDL cholesterol. Our results indicate that an increasing number of the G allele was associated with greater decreases in total cholesterol and LDL cholesterol in response to the low-fat diet, whereas an opposite effect was found in the high-fat diet group.

The disruption of the core molecular clock results in abnormal lipid metabolism, including altered fat storage and lipid transport, and deficits in absorption of dietary lipids (479,490). In this context, it has been demonstrated that melatonin treatment can improve dyslipidemia in both, animal and human studies (491). In humans, daily administration of melatonin for 2 months significantly improved LDL cholesterol among subjects with metabolic syndrome features (492). Moreover, treatment with melatonin and zinc decreased the levels of triglycerides, total cholesterol, LDL cholesterol and increased the levels of HDL cholesterol in type 2 diabetic patients poorly controlled with metformin (493). Despite the fact that the effect of melatonin on lipid profiles has been widely investigated, the mechanisms by which the MTNR1B rs10830963 affects lipid metabolism remains mainly unknown. However, the MTNR1B genetic variant has been associated with melatonin levels and melatonin signaling (478,482). In addition, the MTNR1B rs10830963 polymorphism was significantly related to VLDL and triglyceride levels in Mexican American families (494).

Interestingly, in the present study dietary fat intake modified the effect of the MTNR1B rs10830963 genetic variant on changes in total cholesterol and LDL cholesterol. These results are consistent with a recent meta-analysis that observed significant interactions between the MTNR1B genotype and fat intake (MUFA) on HDL cholesterol levels (188). Although the mechanisms underlying the observed MTNR1B rs10830963 gene-dietary fat interaction need clarification, there is evidence that a high-fat diet could alter the expression and the rhythmic mRNA expression levels of circadian-clock genes and circadian clock-controlled lipogenic genes (485,486,495). For example, Sun et al. (2015) observed a rhythmic expression of the clock- controlled output gene Ppar-a and downstream lipid metabolism genes (Srebp-1c, Fas and Acc1) in normally fed mice (495). Meanwhile, when mice were fed with a high-fat diet, the rhythmic expression in the liver of such genes was significantly altered. Moreover, the genetic effects followed opposite trends according to dietary fat intake on changes in lipid metabolism traits as in our trial. The results are in line with the “differential susceptibility hypothesis”, which proposes that vulnerability genes or risk alleles may function like plasticity genes since genetic risk can be modified by environmental exposures, including dietary factors (496–499). In other words, some individuals might be more responsive to environmental influences in a “for-better-and-for-worse” manner because of the genetic background (497). Consistent with

342 General discussion this hypothesis, we observed that carriers of the risk allele might function as either a protective or a detrimental factor, depending on the differences in dietary fat intake. Epidemiological and functional studies suggest that MTNR1B may play a role in regulating lipid metabolism depending on dietary fat. Nevertheless, further studies are needed to verify this hypothesis.

On the other hand, metabolomics studies have shown that elevated circulating amino acids, particularly branched chain (BCAAs) and aromatic (AAAs) amino acids, might be associated with obesity, insulin resistance and higher risk of type 2 diabetes (500–503). However, it remains unknown whether BCAAs and AAAs are causally implicated in metabolic disturbances (504,505). Several GWAS have revealed common genetic variants determining plasma amino acid levels (504,506–511). Among them, in the present work, a significant association of the PPM1K rs1440581 polymorphism with changes in glucose levels in a large sample of obese European subjects (NUGENOB study) was demonstrated. Furthermore, significant interactions between dietary fat/carbohydrate intake and PPM1K genotype for changes in insulin resistance and b cell function markers were found.

The PPM1K gene encodes mitochondrial protein phosphatase 1K, an activator of the mitochondrial branched-chain alpha-ketoacid dehydrogenase (BCKD) complex, which is a major determinant of the rate of BCAA catabolism (512,513). On one hand, the PPM1K rs1440581 genetic variant has been associated with higher valine and leucine levels, among other amino acids (504,507,509) as well as with the Fischer’s ratio of BCAA to AAA (characteristic of liver fibrosis) (507,514). The same polymorphism was defined by a systems genetics approach as a genetic variant involving susceptibility to type 2 diabetes (515). In fact, the presence of such a polymorphism has also been linked to an increased risk of type 2 diabetes, with an OR of 1.04 per risk allele; 95% CI 1.02,1.07) (504). On the other hand, subjects with mutations of PPM1K (516) and the knock-out Ppmk1 mice (517) presented impaired BCKD activity and high levels of BCAAs and branched chain alpha-ketoacids. Moreover, fat transplant led to a reduction in plasma BCAAs in Ppm1K knock-out mice (518). In this context, Lotta et al. (2016) observed an association between the metabolites accumulated upstream of BCKD action and incidence of T2D, whereas the associations of metabolites accumulated downstream of BCKD action were inconsistent (504). The authors suggested that reduced BCKD activity could be one of the mechanistic links between BCAA metabolism and the risk of type 2 diabetes. However, it should be highlighted that PPM1K could have pleiotropic effects thus limiting the interpretation that the association with glucose metabolism traits is mediated through the BCAAs metabolism and suggesting that other

343 General discussion metabolic pathways could be involved. To our knowledge, despite the evidence about the involvement of PPM1K on BCAAs metabolism and type 2 diabetes, there is paucity of information on the functional effect of the rs1440581 on PPM1K gene expression and consequently on BCKD activity. However, since the common PPM1K genetic variant is located in an intronic region, the splicing process would be altered and with it the structure and function of the protein.

The effect of the PPM1K rs1440581 genetic variant on changes in insulin resistance and b cell function markers was significantly modified by dietary fat/carbohydrate intake. Carriers of the T allele responded better to a high-fat diet than a low-fat diet in terms of insulin, HOMA-IR and HOMA-B. Interestingly, our results are in accordance with the study by Xu et al. (2013) who observed that carriers of the T allele responded better to a high-fat diet than a low-fat diet in terms of body weight, insulin levels and insulin resistance index (252). Although the mechanisms underlying the effects of the modulation of dietary fat/carbohydrate intake on the PPM1K rs1440581 are still poorly understood, a synergistic interference of BCAAs and lipids with the development of insulin resistance has been proposed (519,520). This hypothesis is based on the results in animal studies that have shown that a high-fat diet strengthened the effect of BCAAs supplementation on insulin resistance (520). Although the concentrations of plasma BCAAs were not altered in rats fed a high-fat diet, Kadota et al. (2013) demonstrated that the hepatic BCKD complex was up-regulated in such a model (521). In addition, a glucose challenge resulted in a reduction of circulating BCAA levels and an increased expression of PPM1K in muscle biopsies in normoglycaemic subjects, in agreement with results in animals (522–524). Further experimental studies are needed to understand the mechanisms underlying our results.

Regarding blood pressure levels, candidate gene studies and GWAS have identified a large number of polymorphisms associated with blood pressure or hypertension, which are located in or near genes involved in the renin-angiotensin-aldosterone system, related to enzymes and receptors of the mineral and glucocorticoid pathways and associated with proteins implicated in the structure and or regulation of vascular tone (525). Among them, in the last few years, increased attention has been paid to the NOS3 gene since it is directly involved in BP regulation through nitric oxide (NO) production, which has vasodilator effects (i.e. inhibiting vascular smooth muscle contraction) (526). Interestingly, in the present work and in accordance with other studies, an association between NOS3 rs1799983 genetic variant and an increased risk of hypertension and high to normal DBP was found (527,528). In the rs1799983 polymorphism, a guanine/thymine substitution at exon seven leads to a glutamate/aspartate

344 General discussion substitution at position 298. Several mechanistic studies have been carried out since this genetic variant alters the primary structure of the protein and could alter one or more functional properties of the enzyme (529–534). Different reports have revealed that in the presence of a T instead of a G at nucleotide position 894, NOS3 encodes a protein which leads to a higher susceptibility to cleavage into a 100-kDa fragment (529,530). Thus, the cleaved fragment could decrease NOS3 activity. Nevertheless, other studies concluded that this finding might be a technical artifact (532,533). On the other hand, it has been confirmed that T allele carriers have less NOS3 bound to caveolin-1, which is a protein essential for its activation and therefore to endothelial cell NO production (534).

However, other authors failed to confirm such relationship may be due to the presence of gene-environment interactions. In fact, we found an interaction between the NOS3 rs1799983 polymorphism and fat intake, specifically with SFA and MUFA, and the effect on DBP levels. Although this outcome diverges from the results of Kingah et al. (2010), they are in accordance with the results of Pereira et al. (2007) that suggested a possible gene-diet interaction between the NOS3 genetic variant and diet rich in SFA and cholesterol to influence blood pressure levels (535,536). The mechanisms by which dietary fat modulates the effect of NOS3 genetic variant on BP levels are unknown and can at best only be speculated. In a molecular model a diet rich in olive oil (high MUFA diet) increased the expression of the NOS3 enzyme, whereas a diet rich in SFA decreased the expression of NOS3 (537). Thus, we hypothesized that carriers of the T allele presented lower effect to dietary fat due to a reduced function of NOS3. To establish the mechanism by which dietary fat intake modifies blood pressure depending on NOS3 genotype more studies are needed.

Consistent with a previous research, a significant interaction between NOS3 variant and BMI status on DBP levels was identified (538). This result is in accordance with the initial epidemiological studies suggesting a relationship between obesity and blood pressure levels which has been later supported with the understanding of potential mechanisms involved in both conditions, such as vascular and systemic insulin resistance, dysfunction of the sympathetic nervous system and the renin-angiotensin-aldosterone system, among other pathogenic factors (377,539). On the other hand, NOS3 protein content was shown to be significantly lower in overweight and obese people being inversely associated with body fat mass (540,541). Moreover, oxidative stress, a recognized process implicated in obesity, has been related to NOS3 response, reduced NO production and increased production of reactive oxygen species that may contribute to hypertension in obese rats (542). Therefore, such

345 General discussion scientific evidence suggests that our results may be a genuine interaction rather than a random chance finding.

5. STRENGTHS AND LIMITATIONS

The present research has successfully identified novel GRS or gene phenotype associations as well as GRS or gene-diet interactions on BMI, weight loss and obesity related traits such as lipid metabolism, glucose metabolism and blood pressure. However, some limitations should be declared.

First, each cohort includes subjects of a given ancestry, so the findings may not be extended to other ethnic groups. Although the outcomes related to the PPM1K rs1440581 genetic variant and changes in glucose metabolism traits depending on the genotype were previously published by Xu et al. (2013) in the POUNDS Lost trial (252), further studies are required to generalize the rest of our findings to other ethnic groups.

Second, the relatively small sample size of some of the cohorts might limit the power to detect the effect of the polymorphisms across the obesity and related disorders phenotypes and gene-diet interactions. The lack of large samples could lead to increase the risk of type II errors (failing to detect real differences), mostly if several adjustments are performed (543). For this reason, confounders were limited in the analyses to discover new facts meaning the acceptance of more type I errors (asserting something that is absent) in order to avoid type II error (544). Our approach is consistent with previous investigations that reported that when it is important to discover new facts, we may be willing to accept more type I errors in order to avoid type II errors (544,545). However, the fact that consistent associations and gene-diet interactions were found suggests that potential type II errors were overcome.

Third, the use of self-reported questionnaires in the dietary assessment may involve a possibility of measurement error leading to misclassifications of exposure. On one hand the FGFQ used in the Nutrigenetic Service cohort was validated (397) and registered dietician collect the dietary information.

Fourth, because of the lack of gene expression data we could not confirm the mechanisms involved behind the observed associations between the polymorphisms and phenotypes, and neither how the diet could modify such association. Therefore, in the present study we could only speculate about the underlying biological processes or based on the evidence reported by other authors. For example, the presence of the MTNR1B rs10182181 genotype modified the MTNR1B mRNA expression in human pancreatic islets (478). In addition, the ADCY3 rs713586

346 General discussion polymorphism which is in high linkage disequilibrium with the rs10182181 polymorphism has been related to ADCY3 gene expression (70).

Fifth, circulating levels of some mediators (e.g. NO, melatonin or BCAAs) were not measured, which limited our ability to explore plausible underlying mechanisms. Nonetheless, according to the Mendelian randomization principle, a genetic variant could be a surrogate for the biomarker in causal inference, because it is less likely to be affected by confounding and reverse causation (546).

Finally, another weakness should be acknowledged in the weight-loss dietary intervention studies. In the POUNDS Lost trial and the NUGENOB trial, because low-fat diets are characterized by high-carbohydrate intake and vice versa, to preserve energy balance, it is difficult to define which macronutrient would best explain the observed gene-diet interactions. The same limitation should be acknowledged in the OBEKIT study in which both weight loss diets differed in the content of carbohydrates, protein and fat.

6. COROLLARY

Altogether the results presented in this thesis clearly show that dietary and genetic factors could modulate susceptibility to obesity and its metabolic disorders. Notably, individuals with a high genetic predisposition to obesity, defined by a GRS based on 16 obesity and lipid metabolism related polymorphisms, showed higher adiposity measures and obesity risk than those individuals with a low genetic predisposition. Although scientific evidence suggests the heritability component of obesity, in the present study we observed that genetic factors explain a small percentage of BMI variation as has been found by other authors. However, when we include in the regression model phenotypic features such as age, physical activity and energy intake the percentage of BMI explanation increased. These results confirm that obesity should be treated as a multifactorial disease in which a large number of phenotypic and genotypic factors are involved.

Furthermore, our research work also contributes to better understand not only the role of genetics on body weight loss but also how the diet could modify the association between a polymorphism and body weight regulation. Specifically, for the first time we reported the interaction between the ADCY3 rs10182181 genetic variant and dietary macronutrient distribution on changes in anthropometric and body composition measurements. In addition, the MTNR1B rs101830963 genetic variant interacted with dietary fat intake in response to a hypocaloric diet in terms of body composition and lipid metabolism traits. On the other hand,

347 General discussion the fat/carbohydrate intake modified the association between the PPM1K rs1440581 variant and changes in insulin and insulin resistance traits after a dietary intervention to induced weight loss. Finally, the association between the NOS3 rs1799983 polymorphism and DBP and risk of hypertension was modified by MUFA and PUFA intake, and BMI, respectively.

Nutrigenetics emerges as a good option to further investigate the interindividual susceptibility to metabolic disorders and the response to dietary interventions, and therefore provide personalized nutrition based on the genetic make-up for preventing and treating obesity and its related comorbidities.

348

CONCLUSIONS

Conclusions

1. A large number of single nucleotide polymorphisms has been reported to be related to body fat distribution and body weight regulation, and different studies have described that dietary macronutrient distribution could modify such association. In addition, several epigenetic and metagenomic markers have been identified associated with the obesity phenotype. 2. The proposed food groups frequency questionnaire, based on an exchange system of 19 food groups, has obtained comparable results to other similar models for assessing of energy and macronutrient distribution. 3. Those subjects who were classified as high genetic risk group (genetic risk score >7 risk alleles) presented higher values of adiposity measurements as well as higher risk of obesity and central obesity, than subjects of the low genetic risk group (genetic risk score ≤7 risk alleles). Energy, total protein, vegetable protein, saturated fatty acids, polyunsaturated fatty acids and total carbohydrates intake modified the association between the genetic risk score and percentage of body fat mass and obesity risk. 4. The statistical model selection technique, least angle regression analysis, could help to implement new criteria for the identification of BMI predictors since obesity is a multifactorial disease in which a large number of phenotypic and genotypic features are involved. 5. Least angle regression analysis identified as significant predictors of BMI age, energy intake, physical activity and polymorphisms located near or in FTO (rs9939609), APOE (rs429358), PPARG (rs181282) and PPARA (rs1800206). 6. The ADCY3 rs10182181 genetic variant interacted with dietary macronutrient distribution on changes in weight, waist circumference, fat mass, percentage of fat mass, percentage of lean mass, trunk fat, android fat, gynoid fat and visceral fat after a 16-week weight loss dietary intervention. Carriers of the minor allele of the ADCY3 variant might have a better response to a weight loss dietary intervention by choosing a low-fat diet than a moderately-high-protein diet. 7. The circadian rhythms related MTNR1B rs10830963 polymorphism was associated with individual differences in weight loss induced by a hypocaloric diet among women. This association was influenced by FTO rs9939609 and MC4R rs17782313 loci and modified by baseline protein intake. 8. Fat intake modified the effect of the MTNR1B rs10830963 genetic variant on changes in body fatness and composition. Among individuals with the G allele consuming the low-fat diet showed greater effect on changes in weight, BMI, waist circumference and total body fat, compared with the high-fat diet over the 6-month dietary intervention.

351 Conclusions

9. The MTNR1B rs10830963 variant interacted with dietary fat intake on changes in total cholesterol and LDL cholesterol after a 2-year diet-induced weight loss. Carriers of the G allele may benefit more in the improvement of their lipid profile by consuming a low-fat diet instead of a high-fat diet. Conversely, for improving lipid metabolism markers, TT individuals may benefit more from a high-fat diet. 10. The PPM1K rs1440581 polymorphism was associated with changes in glucose levels after 10-week dietary intervention. In addition, dietary fat/carbohydrate intake interacted with PPM1K genotypes for changes in insulin resistance and b cell function markers. Individuals with the CC genotype might be more responsive to a low-fat/high-carbohydrate diet in lowering insulin, HOMA-IR and HOMA-B compared with those without this genotype. 11. NOS3 rs1799983 TT genotype subjects showed a higher predisposition of hypertension. Moreover, carriers of the T allele presented greater values of DBP. This association was influenced by dietary fat intake (saturated fatty acids and monounsaturated fatty acids) and BMI status.

352

REFERENCES

References

1. World Health Organization (WHO). Obesity and overweight. 2016. 2. Tchernof A, Després J-P. Pathophysiology of human visceral obesity: an update. Physiol Rev. 2013;93:359– 404. 3. Abete I, Astrup A, Martínez JA, Thorsdottir I, Zulet MA. Obesity and the metabolic syndrome: role of different dietary macronutrient distribution patterns and specific nutritional components on weight loss and maintenance. Nutr Rev. 2010;68:214–31. 4. Forouzanfar MH, Alexander L, Anderson HR, Bachman VF, Biryukov S, Brauer M, Burnett R, Casey D, Coates MM, Cohen A, et al. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks in 188 countries, 1990-2013: A systematic analysis for the Global Burden of Disease Study 2013. The Lancet. 2015;386:2287–323. 5. Biddle SJH, Bengoechea García E, Pedisic Z, Bennie J, Vergeer I, Wiesner G. Screen time, other sedentary behaviours, and obesity risk in adults: a review of reviews. Curr Obes Rep. 2017;6:134-147. 6. St-Onge M-P. Sleep-obesity relation: underlying mechanisms and consequences for treatment. Obes Rev. 2017;18:34–9. 7. Kim TJ, Roesler NM, von dem Knesebeck O. Causation or selection - examining the relation between education and overweight/obesity in prospective observational studies: a meta-analysis. Obes Rev. 2017;18:660-672. 8. El-Sayed AM, Scarborough P, Galea S. Unevenly distributed: a systematic review of the health literature about socioeconomic inequalities in adult obesity in the United Kingdom. BMC Public Health. 2012;12:18–30. 9. Weaver JU. Classical endocrine diseases causing obesity. Front Horm Res. 2008;36:212–28. 10. Martínez de Morentin-Aldabe BE, Hernández-Ruiz de Eguilaz M, Pérez-Díez S, Martínez JA. Influencia de la administración de psicofármacos en el aumento del peso corporal. Rev Española Nutr Humana y Dietética. 2013;17:17–26. 11. Singh RK, Kumar P, Mahalingam K. Molecular genetics of human obesity: a comprehensive review. C R Biol. 2017;340:87–108. 12. Milagro FI, Mansego ML, De Miguel C, Martínez JA. Dietary factors, epigenetic modifications and obesity outcomes: Progresses and perspectives. Molecular Aspects of Medicine. 2013;34:782–812. 13. Khan MJ, Gerasimidis K, Edwards CA, Shaikh MG. Role of gut microbiota in the aetiology of obesity: proposed mechanisms and review of the literature. Journal of Obesity. 2016;2016:7353642. 14. Franks PW, Pearson E, Florez JC. Gene-environment and gene-treatment interactions in type 2 diabetes: progress, pitfalls, and prospects. Diabetes Care. 2013;36:1413–21. 15. Martinez JA, Navas-Carretero S, Saris WH, Astrup A. Personalized weight loss strategies-the role of macronutrient distribution. Nat Rev. 2014;10:749-60. 16. World Health Organization (WHO). Obesity: preventing and managing the global epidemic. 2000. 17. Salas-Salvadó J, Rubio MA, Barbany M, Moreno B, Aranceta J, Bellido D, Blay V, Carraro R, Xavier F, Foz M, et al. Consenso SEEDO 2007 para la evaluación del sobrepeso y la obesidad y el establecimiento de criterios de intervención terapéutica. Med Clínica. 2007;128:184–96. 18. De Schutter A, Lavie CJ, Kachur S, Patel DA, Milani R V. Body composition and mortality in a large cohort with preserved ejection fraction: untangling the obesity paradox. Mayo Clin Proc. 2014;89:1072–9. 19. Nazare JA, Smith J, Borel AL, Aschner P, Barter P, Van Gaal L, Tan CE, Wittchen HU, Matsuzawa Y, Kadowaki T, et al. Usefulness of measuring both body mass index and waist circumference for the estimation of visceral adiposity and related cardiometabolic risk profile (from the INSPIRE ME IAA Study). Am J Cardiol. 2015;115:307–15. 20. Fields DA, Goran MI, McCrory MA. Body-composition assessment via air-displacement plethysmography in

355 References adults and children: a review. Am J Clin Nutr. 2002;75:453–67. 21. Sun G, French C, Martin G, Al E. Comparison of multifrequency bioelectrical impedance analysis with dual energy x-ray absorptiometry for assessment of percentage body fat in a large, health population. Am J Clin Nutr. 2005;81:74–8. 22. Aranceta J, Pérez Rodrigo C, Serra Majem L, Ribas Barba L, Quiles Izquierdo J, Vioque J, Tur Marí J, Mataix Verdú J, Llopis González J, Tojo R, et al. Consenso SEEDO 2000 para la evaluación del sobrepeso y la obesidad y el establecimiento de criterios de intervención terapéutica. Med Clin (Barc). 2000;115:587–97. 23. Scully T. Society at large. Nature. 2014;508:50–1. 24. Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono C, Mullany EC, Biryukov S, Abbafati C, Abera SF, et al. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980-2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2014;384:766–81. 25. Kelly T, Yang W, Chen C-S, Reynolds K, He J. Global burden of obesity in 2005 and projections to 2030. Int J Obes. 2008;32:1431–7. 26. Neel J V. Diabetes mellitus: a “thrifty” genotype rendered detrimental by “progress”? 1962. Bull World Health Organ. 1999;77:353–62. 27. Baig U, Belsare P, Watve M, Jog M. Can thrifty gene(s) or predictive fetal programming for thriftiness lead to obesity? J Obes. 2011;2011:1–11. 28. Speakman JR. Thrifty genes for obesity, an attractive but flawed idea, and an alternative perspective: the “drifty gene” hypothesis. Int J Obes. 2008;32:1611–7. 29. Speakman JR. Evolutionary perspectives on the obesity epidemic: adaptive, maladaptive, and neutral viewpoints. Annu Rev Nutr. 2013;33:289–317. 30. Tenesa A, Haley CS. The heritability of human disease: estimation, uses and abuses. Nat Rev Genet. 2013;14:139–49. 31. Min J, Chiu DT, Wang Y. Variation in the heritability of body mass index based on diverse twin studies: a systematic review. Obes Rev. 2013;14:871–82. 32. Stunkard AJ, Foch TT, Hrubec Z. A twin study of human obesity. JAMA. 1986;256:51–4. 33. Nan C, Guo B, Warner C, Fowler T, Barrett T, Boomsma D, Nelson T, Whitfield K, Beunen G, Thomis M, et al. Heritability of body mass index in pre-adolescence, young adulthood and late adulthood. Eur J Epidemiol. 2012;27:247–53. 34. Sørensen TI, Price RA, Stunkard AJ, Schulsinger F. Genetics of obesity in adult adoptees and their biological siblings. BMJ. 1989;298:87–90. 35. Knowler WC, Pettitt DJ, Saad MF, Charles MA, Nelson RG, Howard B V, Bogardus C, Bennett PH. Obesity in the Pima Indians: its magnitude and relationship with diabetes. Am J Clin Nutr. 1991;53:1543S–1551S. 36. Farooqi I, Rahilly SO. Monogenic obesity in humans. Annu Rev Med. 2005;56:443–56. 37. Albuquerque D, Stice E, Rodríguez-López R, Manco L, Nóbrega C. Current review of genetics of human obesity: from molecular mechanisms to an evolutionary perspective. Mol Genet Genomics. 2015;290:1191–221. 38. Farooqi SI. Genetic, molecular and physiological mechanisms involved in human obesity: Society for endocrinology medal lecture 2012. Clin Endocrinol. 2015;82:23–8. 39. González-Jiménez E, Aguilar Cordero MJ, Padilla López CA, García García I. Monogenic human obesity: role of the leptin-melanocortin system in the regulation of food intake and body weight in humans. An Sist Sanit Navar. 2012;35:285–93. 40. El-Sayed Moustafa JS, Froguel P. From obesity genetics to the future of personalized obesity therapy. Nat Rev

356 References

Endocrinol. 2013;9:402–13. 41. Razquin C, Marti A, Martínez JA. Evidences on three relevant obesogenes: MC4R, FTO and PPARγ. Approaches for personalized nutrition. Mol Nutr Food Res. 2011;55:136–49. 42. Xia Q, Grant SFA. The genetics of human obesity. Ann N Y Acad Sci. 2013;1281:178–90. 43. Rankinen T, Zuberi A, Chagnon YC, Weisnagel SJ, Argyropoulos G, Walts B, Pérusse L, Bouchard C. The human obesity gene map: the 2005 update. Obesity. 2006;14:529–644. 44. Bender N, Allemann N, Marek D, Vollenweider P, Waeber G, Mooser V, Egger M, Bochud M. Association between variants of the leptin receptor gene (LEPR) and overweight: a systematic review and an analysis of the colaus study. PLoS One. 2011;6:1–14. 45. Peng S, Zhu Y, Xu F, Ren X, Li X, Lai M. FTO gene polymorphisms and obesity risk: a meta-analysis. BMC Med. 2011;9:71–86. 46. Wang R, Zhou D, Xi B, Ge X, Zhu P, Wang B, Zhou M, Huang Y, Liu J, Yu Y, et al. ENPP1/PC-1 gene K121Q polymorphism is associated with obesity in European adult populations: evidence from a meta-analysis involving 24,324 subjects. Biomed Environ Sci. 2011;24:200–6. 47. Zhao T, Zhao J, Lv J, Nzekebaloudou M. Meta-analysis on the effect of the Ala54Thr polymorphism of the fatty acid-binding protein 2 gene on body mass index. Nutr Metab Cardiovasc Dis. 2011;21:823–9. 48. Xi B, Chandak GR, Shen Y, Wang Q, Zhou D. Association between common polymorphism near the MC4R gene and obesity risk: a systematic review and meta-analysis. PLoS One. 2012;7:e45731. 49. Yu Z, Han S, Cao X, Zhu C, Wang X, Guo X. Genetic polymorphisms in adipokine genes and the risk of obesity: a systematic review and meta-analysis. Obesity. 2012;20:396–406. 50. Galbete C, Toledo E, Martínez-González MA, Martínez JA, Guillén-Grima F, Marti A. Pro12Ala variant of the PPARG2 gene increases body mass index: an updated meta-analysis encompassing 49,092 subjects. Obesity. 2013;21:1486–95. 51. Brondani LA, Assmann TS, De Souza BM, Bouças AP, Canani LH, Crispim D. Meta-analysis reveals the association of common variants in the Uncoupling Protein (UCP) 1-3 genes with body mass index variability. PLoS One. 2014;9:e96411. 52. Stijnen P, Tuand K, Varga T V., Franks PW, Aertgeerts B, Creemers JWM. The association of common variants in PCSK1 with obesity: a HuGE review and meta-analysis. Am J Epidemiol. 2014;180:1051–65. 53. Zhang H, Wu J, Yu L. Association of Gln27Glu and Arg16Gly polymorphisms in beta2-adrenergic receptor gene with obesity susceptibility: A meta-analysis. PLoS One. 2014;9:e100489. 54. Zhang L, Yuan LH, Xiao Y, Lu MY, Zhang LJ, Wang Y. Association of leptin gene -2548 G/A polymorphism with obesity: a meta-analysis. Ann Nutr Metab. 2014;64:127–36. 55. Zhang M, Wang M, Zhao Z-T. Uncoupling protein 2 gene polymorphisms in association with overweight and obesity susceptibility: a meta-analysis. Meta Gene. 2014;2:143–59. 56. Mansoori A, Amini M, Kolahdooz F, Seyedrezazadeh E. Obesity and Pro12Ala polymorphism of peroxisome proliferator-activated receptor-gamma gene in healthy adults: a systematic review and meta-analysis. Ann Nutr Metab. 2015;67:104–18. 57. Yan J, Wang X, Tao H, Yang W, Luo M, Lin F. Lack of association between leptin G-2548A polymorphisms and obesity risk: evidence based on a meta-analysis. Obes Res Clin Pract. 2015;9:389–97. 58. Yao Y shui, Li J, Jin Y long, Chen Y, He L ping. Association between PPAR-gamma Pro12Ala polymorphism and obesity: a meta-analysis. Mol Biol Rep. 2015;42:1029–38. 59. Li HL, Zhang YJ, Chen XP, Luo JQ, Liu SY, Zhang ZL. Association between GNB3 c.825C>T polymorphism and the

357 References risk of overweight and obesity: a meta-analysis. Meta Gene. 2016;9:18–25. 60. McPherson JD, Marra M, Hillier L, Waterston RH, Chinwalla A, Wallis J, Sekhon M, Wylie K, Mardis ER, Wilson RK, et al. A physical map of the human genome. Nature. 2001;409:934–41. 61. Gibbs RA, Belmont JW, Hardenbol P, Willis TD, Yu F, Yang H, Ch’ang L-Y, Huang W, Liu B, et al. The International HapMap Project. Nature. 2003;426:789–96. 62. Larder R, Lim CT, Coll AP. Genetic aspects of human obesity. Handbook of clinical neurology. 2014. p93–106. 63. Herbert A, Gerry NP, McQueen MB, Heid IM, Pfeufer A, Illig T, Wichmann H-E, Meitinger T, Hunter D, Hu FB, et al. A common genetic variant is associated with adult and childhood obesity. Science. 2006;312:279–83. 64. Dina C, Meyre D, Gallina S, Durand E, Körner A, Jacobson P, Carlsson LMS, Kiess W, Vatin V, Lecoeur C, et al. Variation in FTO contributes to childhood obesity and severe adult obesity. Nat Genet. 2007;39:724–6. 65. Frayling TM, Timpson NJ, Weedon MN, Zeggini E, Freathy RM, Lindgren CM, Perry JRB, Elliott KS, Lango H, Rayner NW, et al. A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science. 2007;316:889–94. 66. Scuteri A, Sanna S, Chen WM, Uda M, Albai G, Strait J, Najjar S, Nagaraja R, Orru M, Usala G, et al. Genome- wide association scan shows genetic variants in the FTO gene are associated with obesity-related traits. PLoS Genet. 2007;3:1200–10. 67. Loos RJF, Lindgren CM, Li S, Wheeler E, Zhao JH, Prokopenko I, Inouye M, Freathy RM, Attwood AP, Beckmann JS, et al. Common variants near MC4R are associated with fat mass, weight and risk of obesity. Nat Genet. 2008;40:768–75. 68. Willer CJ, Speliotes EK, Loos RJF, Li SX, Lindgren CM, Heid IM, Berndt SI, Elliott AL, Jackson AU, Lamina C, et al. Six new loci associated with body mass index highlight a neuronal influence on body weight regulation. Nat Genet. 2009;41:25–34. 69. Thorleifsson G, Walters GB, Gudbjartsson DF, Steinthorsdottir V, Sulem P, Helgadottir A, Styrkarsdottir U, Gretarsdottir S, Thorlacius S, Jonsdottir I, et al. Genome-wide association yields new sequence variants at seven loci that associate with measures of obesity. Nat Genet. 2009;41:18–24. 70. Speliotes EK, Willer CJ, Berndt SI, Monda KL, Thorleifsson G, Jackson AU, Lango Allen H, Lindgren CM, Luan J, Mägi R, et al. Association analyses of 249,796 individuals reveal 18 new loci associated with body mass index. Nat Genet. 2010;42:937–48. 71. Berndt SI, Gustafsson S, Mägi R, Ganna A, Wheeler E, Feitosa MF, Justice AE, Monda KL, Croteau-Chonka DC, Day FR, et al. Genome-wide meta-analysis identifies 11 new loci for anthropometric traits and provides insights into genetic architecture. Nat Genet. 2013;45:501–12. 72. Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Day FR, Powell C, Vedantam S, Buchkovich ML, Yang J, et al. Genetic studies of body mass index yield new insights for obesity biology. Nature. 2015;518:197–206. 73. Monda KL, Chen GK, Taylor KC, Palmer C, Edwards TL, Lange LA, Ng MCY, Adeyemo AA, Allison MA, Bielak LF, et al. A meta-analysis identifies new loci associated with body mass index in individuals of African ancestry. Nat Genet. 2013;45:690–6. 74. Wen W, Zheng W, Okada Y, Takeuchi F, Tabara Y, Hwang JY, Dorajoo R, Li H, Tsai FJ, Yang X, et al. Meta- analysis of genome-wide association studies in East Asian-ancestry populations identifies four new loci for body mass index. Hum Mol Genet. 2014;23:5492–504. 75. Fox CS, Heard-Costa N, Cupples LA, Dupuis J, Vasan RS, Atwood LD. Genome-wide association to body mass index and waist circumference: the Framingham Heart Study 100K project. BMC Med Genet. 2007;8:S18–25. 76. Hinney A, Nguyen TT, Scherag A, Friedel S, Brönner G, Müller TD, Grallert H, Illig T, Wichmann HE, Rief W, et

358 References al. Genome wide association (GWA) study for early onset extreme obesity supports the role of fat mass and obesity associated gene (FTO) variants. PLoS One. 2007;2:e1361. 77. Liu YJ, Liu XG, Wang L, Dina C, Yan H, Liu JF, Levy S, Papasian CJ, Drees BM, Hamilton JJ, et al. Genome-wide association scans identified CTNNBL1 as a novel gene for obesity. Hum Mol Genet. 2008;17:1803–13. 78. Cotsapas C, Speliotes EK, Hatoum IJ, Greenawalt DM, Dobrin R, Lum PY, Suver C, Chudin E, Kemp D, Reitman M, et al. Common body mass index-associated variants confer risk of extreme obesity. Hum Mol Genet. 2009;18:3502–7. 79. Lowe JK, Maller JB, Pe’er I, Neale BM, Salit J, Kenny EE, Shea JL, Burkhardt R, Smith JG, Ji W, et al. Genome- wide association studies in an isolated founder population from the Pacific Island of Kosrae. PLoS Genet. 2009;5:e1000365. 80. Meyre D, Delplanque J, Chèvre J-C, Lecoeur C, Lobbens S, Gallina S, Durand E, Vatin V, Degraeve F, Proença C, et al. Genome-wide association study for early-onset and morbid adult obesity identifies three new risk loci in European populations. Nat Genet. 2009;41:157–9. 81. Norris JM, Langefeld CD, Talbert ME, Wing MR, Haritunians T, Fingerlin TE, Hanley AJG, Ziegler JT, Taylor KD, Haffner SM, et al. Genome-wide association study and follow-up analysis of adiposity traits in Hispanic Americans: the IRAS Family Study. Obesity. 2009;17:1932–41. 82. Polašek O, Marušić A, Rotim K, Hayward C, Vitart V, Huffman J, Campbell S, Janković S, Boban M, Biloglav Z, et al. Genome-wide association study of anthropometric traits in Korčula Island, Croatia. Croat Med J. 2009;50:7–16. 83. Johansson A, Marroni F, Hayward C, Franklin CS, Kirichenko A V, Jonasson I, Hicks A a, Vitart V, Isaacs A, Axenovich T, et al. Linkage and genome-wide association analysis of obesity-related phenotypes: association of weight with the MGAT1 gene. Obesity. 2010;18:803–8. 84. Liu JZ, Medland SE, Wright MJ, Heath AC, Pamela AF, Duncan A, Montgomery GW, Martin NG, Allan F. Genome-wide association study of height and body mass index in Australian twin families. Twin Res Hum Genet. 2010;13:179–93. 85. Scherag A, Dina C, Hinney A, Vatin V, Scherag S, Vogel CIG, Müller TD, Grallert H, Wichmann HE, Balkau B, et al. Two new loci for body-weight regulation identified in a joint analysis of genome-wide association studies for early-onset extreme obesity in French and German study groups. PLoS Genet. 2010;6:2–11. 86. Tönjes A, Zeggini E, Kovacs P, Böttcher Y, Schleinitz D, Dietrich K, Morris AP, Enigk B, Rayner NW, Koriath M, et al. Association of FTO variants with BMI and fat mass in the self-contained population of Sorbs in Germany. Eur J Hum Genet. 2010;18:104–10. 87. Benjamin AM, Suchindran S, Pearce K, Rowell J, Lien LF, Guyton JR, McCarthy JJ. Gene by sex interaction for measures of obesity in the Framingham Heart Study. J Obes. 2011;2011:329038–46. 88. Croteau-Chonka DC, Marvelle AF, Lange EM, Lee NR, Adair LS, Lange LA, Mohlke KL. Genome-wide association study of anthropometric traits and evidence of interactions with age and study year in Filipino women. Obesity (Silver Spring). 2011;19:1019–27. 89. Jiao H, Arner P, Hoffstedt J, Brodin D, Dubern B, Czernichow S, van’t Hooft F, Axelsson T, Pedersen O, Hansen T, et al. Genome wide association study identifies KCNMA1 contributing to human obesity. BMC Med Genomics. 2011;4:51–61. 90. Malhotra A, Kobes S, Knowler WC, Baier LJ, Bogardus C, Hanson RL. A genome-wide association study of BMI in American Indians. Obesity. 2011;19:2102–6. 91. Paternoster L, Evans DM, Aagaard Nohr E, Holst C, Gaborieau V, Brennan P, Prior Gjesing A, Grarup N, Witte DR, Jørgensen T, et al. Genome-wide population-based association study of extremely overweight young adults -

359 References the GOYA study. PLoS One. 2011;6:e24303. 92. Wang K, Li WD, Zhang CK, Wang Z, Glessner JT, Grant SFA, Zhao H, Hakonarson H, Price RA. A genome-wide association study on obesity and obesity-related traits. PLoS One. 2011;6:3–8. 93. Bradfield JP, Taal HR, Timpson NJ, Scherag A, Lecoeur C, Warrington NM, Hypponen E, Holst C, Valcarcel B, Thiering E, et al. A genome-wide association meta-analysis identifies new childhood obesity loci. Nat Genet. 2012;44:526–31. 94. Comuzzie AG, Cole SA, Laston SL, Voruganti VS, Haack K, Gibbs RA, Butte NF. Novel genetic loci identified for the pathophysiology of childhood obesity in the Hispanic Population. PLoS One. 2012;7:e51954. 95. Dorajoo R, Blakemore a IF, Sim X, Ong RT-H, Ng DPK, Seielstad M, Wong T-Y, Saw S-M, Froguel P, Liu J, et al. Replication of 13 obesity loci among Singaporean Chinese, Malay and Asian-Indian populations. Int J Obes. 2012;36:159–63. 96. Melka MG, Bernard M, Mahboubi A, Abrahamowicz M, Paterson AD, Syme C, Lourdusamy A, Schumann G, Leonard GT, Perron M, et al. Genome-wide scan for loci of adolescent obesity and their relationship with blood pressure. J Clin Endocrinol Metab. 2012;97:e145–50. 97. Ng MCY, Hester JM, Wing MR, Li J, Xu J, Hicks PJ, Roh BH, Lu L, Divers J, Langefeld CD, et al. Genome-wide association of BMI in African Americans. Obesity (Silver Spring). 2012;20:622–7. 98. Okada Y, Kubo M, Ohmiya H, Takahashi A, Kumasaka N, Hosono N, Maeda S, Wen W, Dorajoo R, Go MJ, et al. Common variants at CDKAL1 and KLF9 are associated with body mass index in east Asian populations. Nat Genet. 2012;44:302–6. 99. Wang KS, Liu X, Zheng S, Zeng M, Pan Y, Callahan K. A novel locus for body mass index on 5p15.2: a meta- analysis of two genome-wide association studies. Gene. 2012;500:80–4. 100. Wen W, Cho YS, Zheng W, Dorajoo R, Kato N, Qi L, Chen C, Delahanty RJ, Okada Y. Meta-analysis identifies common variants associated with body mass index in East Asians. Nat Genet. 2012;44:307–11. 101. Yang J, Loos RJF, Powell JE, Medland SE, Speliotes EK, Chasman DI, Rose LM, Thorleifsson G, Steinthorsdottir V, Mägi R, et al. FTO genotype is associated with phenotypic variability of body mass index. Nature. 2012;490:267– 72. 102. Graff M, Ngwa JS, Workalemahu T, Homuth G, Schipf S, Teumer A, Völzke H, Wallaschofski H, Abecasis GR, Edward L, et al. Genome-wide analysis of BMI in adolescents and young adults reveals additional insight into the effects of genetic loci over the life course. Hum Mol Genet. 2013;22:3597–607. 103. Kim HJ, Yoo YJ, Ju YS, Lee S, Cho S Il, Sung J, Kim J Il, Seo JS. Combined linkage and association analyses identify a novel locus for obesity near PROX1 in Asians. Obesity. 2013;21:2405–12. 104. Melén E, Granell R, Kogevinas M, Strachan D, Gonzalez JR, Wjst M, Jarvis D, Ege M, Braun-Fahrländer C, Genuneit J, et al. Genome-wide association study of body mass index in 23000 individuals with and without asthma. Clin Exp Allergy. 2013;43:463–74. 105. Namjou B, Keddache M, Marsolo K, Wagner M, Lingren T, Cobb B, Perry C, Kennebeck S, Holm IA, Li R, et al. EMR-linked GWAS study: investigation of variation landscape of loci for body mass index in children. Front Genet. 2013;4:1–9. 106. Plourde M, Vohl M-C, Bellis C, Carless M, Dyer T, Dolley G, Marette A, Després J-P, Bouchard C, Blangero J, et al. A variant in the LRRFIP1 gene is associated with adiposity and inflammation. Obesity. 2013;21:185–92. 107. Wheeler E, Huang N, Bochukova EG, Keogh JM, Lindsay S, Garg S, Henning E, Blackburn H, Loos RJF, Wareham NJ, et al. Genome-wide SNP and CNV analysis identifies common and low-frequency variants associated with severe early-onset obesity. Nat Genet. 2013;45:513–7.

360 References

108. Pei YF, Zhang L, Yongjun Liu, Li J, Shen H, Liu YZ, Tian Q, He H, Wu S, Ran S, et al. Meta-analysis of genome- wide association data identifies novel susceptibility loci for obesity. Hum Mol Genet. 2014;23:820–30. 109. Scannell Bryan M, Argos M, Pierce B, Tong L, Rakibuz-Zaman M, Ahmed A, Rahman M, Islam T, Yunus M, Parvez F, et al. Genome-wide association studies and heritability estimates of body mass index related phenotypes in Bangladeshi adults. PLoS One. 2014;9:e105062. 110. Yang F, Chen X, Ta L, Shen J, Li D, Zhang F, BY S, Deng H. Genome wide association study: searching for genes underlying body mass index in Chinese. Biomed Env Sci. 2014;27:360–70. 111. Anderson D, Cordell HJ, Fakiola M, Francis RW, Syn G, Scaman ESH, Davis E, Miles SJ, McLeay T, Jamieson SE, et al. First genome-wide association study in an Australian Aboriginal population provides insights into genetic risk factors for body mass index and type 2 diabetes. PLoS One. 2015;10:e0119333. 112. McQueen MB, Boardman JD, Domingue BW, Smolen A, Tabor J, Killeya-Jones L, Halpern CT, Whitsel EA, Harris KM. The National Longitudinal Study of Adolescent to Adult Health (Add Health) Sibling Pairs Genome-Wide Data. Behav Genet. 2014;45:12–23. 113. Warrington NM, Howe LD, Paternoster L, Kaakinen M, Herrala S, Huikari V, Wu YY, Kemp JP, Timpson NJ, Pourcain BS, et al. A genome-wide association study of body mass index across early life and childhood. Int J Epidemiol. 2015;44:700–12. 114. Felix JF, Bradfield JP, Monnereau C, Van Der Valk RJP, Stergiakouli E, Chesi A, Gaillard R, Feenstra B, Thiering E, Kreiner-M??ller E, et al. Genome-wide association analysis identifies three new susceptibility loci for childhood body mass index. Hum Mol Genet. 2016;25:389–403. 115. Hwang J-Y, Lee HJ, Go MJ, Jang HB, Park SI, Kim B-J, Lee H-J. An integrative study identifies KCNC2 as a novel predisposing factor for childhood obesity and the risk of diabetes in the Korean population. Sci Rep. 2016;14:33043–55. 116. Lee M, Kwon DY, Kim MS, Choi CR, Park MY, Kim AJ. Genome-wide association study for the interaction between BMR and BMI in obese Korean women including overweight. Nutr Res Pract. 2016;10:115–24. 117. Ried JS, Jeff M. J, Chu AY, Bragg-Gresham JL, van Dongen J, Huffman JE, Ahluwalia TS, Cadby G, Eklund N, Eriksson JJG, et al. A principal component meta-analysis on multiple anthropometric traits identifies novel loci for body shape. Nat Commun. 2016;7:1–11. 118. Salinas YD, Wang L, Dewan AT. Multiethnic genome-wide association study identifies ethnic-specific associations with body mass index in Hispanics and African Americans. BMC Genet. 2016;17:78–91. 119. Justice AE, Winkler TW, Feitosa MF, Graff M, Fisher VA, Young K, Barata L, Deng X, Czajkowski J, Hadley D, et al. Genome-wide meta-analysis of 241,258 adults accounting for smoking behaviour identifies novel loci for obesity traits. Nat Commun. 2017;8:1497714996. 120. Müller M, Laperpusch M, Ender J, Schautz B, Heller M, Bosy-Westphal A. Beyond the body mass index: tracking body composition in the pathogenesis of obesity and the metabolic syndrome. Obes Rev. 2012;13:6–13. 121. Goni L, Milagro FI, Cuervo M, Martínez JA. Single-nucleotide polymorphisms and DNA methylation markers associated with central obesity. Nutr Rev. 2014;72:673–90. 122. Chambers JC, Elliott P, Zabaneh D, Zhang W, Li Y, Froguel P, Balding D, Scott J, Kooner JS. Common genetic variation near MC4R is associated with waist circumference and insulin resistance. Nat Genet. 2008;40:716–8. 123. Heard-Costa NL, Zillikens MC, Monda KL, Johansson A, Harris TB, Fu M, Haritunians T, Feitosa MF, Aspelund T, Eiriksdottir G, et al. NRXN3 is a novel locus for waist circumference: a genome-wide association study from the CHARGE Consortium. PLoS Genet. 2009;5:e1000539. 124. Lindgren CM, Heid IM, Randall JC, Lamina C, Steinthorsdottir V, Qi L, Speliotes EK, Thorleifsson G, Willer CJ,

361 References

Herrera BM, et al. Genome-wide association scan meta-analysis identifies three Loci influencing adiposity and fat distribution. PLoS Genet. 2009;5:e1000508. 125. Fall T, Ingelsson E. Genome-wide association studies of obesity and metabolic syndrome. Mol Cell Endocrinol. 2014;382:740–57. 126. Heid IM, Jackson AU, Randall JC, Winkler TW, Qi L, Steinthorsdottir V, Thorleifsson G, Zillikens MC, Speliotes EK, Magi R, et al. Meta-analysis identifies 13 new loci associated with waist-hip ratio and reveals sexual dimorphism in the genetic basis of fat distribution. Nat Genet. 2010;42:949–60. 127. Liu CT, Monda KL, Taylor KC, Lange L, Demerath EW, Palmas W, Wojczynski MK, Ellis JC, Vitolins MZ, Liu S, et al. Genome-wide association of body fat distribution in african ancestry populations suggests new Loci. PLoS Genet. 2013;9:e1003681. 128. Randall JC, Winkler TW, Kutalik Z, Berndt SI, Jackson AU, Monda KL, Kilpelainen TO, Esko T, Magi R, Li S, et al. Sex-stratified genome-wide association studies including 270,000 individuals show sexual dimorphism in genetic loci for anthropometric traits. PLoS Genet. 2013;9:e1003500. 129. Shungin D, Winkler TW, Croteau-Chonka DC, Ferreira T, Locke AE, Magi R, Strawbridge RJ, Pers TH, Fischer K, Justice AE, et al. New genetic loci link adipose and insulin biology to body fat distribution. Nature. 2015;518:187–96. 130. Wen W, Kato N, Hwang J-Y, Guo X, Tabara Y, Li H, Dorajoo R, Yang X, Tsai F-J, Li S, et al. Genome-wide association studies in East Asians identify new loci for waist-hip ratio and waist circumference. Sci Rep. 2016;6:17958–67. 131. Cho YS, Go MJ, Kim YJ, Heo JY, Oh JH, Ban HJ, Yoon D, Lee MH, Kim DJ, Park M, et al. A large-scale genome- wide association study of Asian populations uncovers genetic factors influencing eight quantitative traits. Nat Genet. 2009;41:527–34. 132. Kilpeläinen TO, Qi L, Brage S, Sharp SJ, Sonestedt E, Demerath E, Ahmad T, Mora S, Kaakinen M, Sandholt CH, et al. Physical activity attenuates the influence of FTO variants on obesity risk: a meta-analysis of 218,166 adults and 19,268 children. PLoS Med. 2011;8:e1001116. 133. Kraja AT, Vaidya D, Pankow JS, Goodarzi MO, Assimes TL, Kullo IJ, Sovio U, Mathias RA, Sun Y V, Franceschini N, et al. A bivariate genome-wide approach to metabolic syndrome: STAMPEED consortium. Diabetes. 2011;60:1329–39. 134. Kristiansson K, Perola M, Tikkanen E, Kettunen J, Surakka I, Havulinna AS, Stancakova A, Barnes C, Widen E, Kajantie E, et al. Genome-wide screen for metabolic syndrome susceptibility Loci reveals strong lipid gene contribution but no evidence for common genetic basis for clustering of metabolic syndrome traits. Circ Genet. 2012;5:242–9. 135. Scott WR, Zhang W, Loh M, Tan ST, Lehne B, Afzal U, Peralta J, Saxena R, Ralhan S, Wander GS, et al. Investigation of genetic variation underlying central obesity amongst South Asians. PLoS One. 2016;11:1–17. 136. Kilpeläinen TO, Zillikens MC, Stančákova A, Finucane FM, Ried JS, Langenberg C, Zhang W, Beckmann JS, Luan J, Vandenput L, et al. Genetic variation near IRS1 associates with reduced adiposity and an impaired metabolic profile. Nat Genet. 2011;43:753–60. 137. Lu Y, Day FR, Gustafsson S, Buchkovich ML, Na J, Bataille V, Cousminer DL, Dastani Z, Drong AW, Esko T, et al. New loci for body fat percentage reveal link between adiposity and cardiometabolic disease risk. Nat Commun. 2016;7:10495–510. 138. Chu AY, Deng X, Fisher VA, Drong A, Zhang Y, Feitosa MF, Liu C-T, Weeks O, Choh AC, Duan Q, et al. Multiethnic genome-wide meta-analysis of ectopic fat depots identifies loci associated with adipocyte development and differentiation. Nat Genet. 2017;49:125–30.

362 References

139. Norris JM, Langefeld CD, Talbert ME, Wing MR, Haritunians T, Fingerlin TE, Hanley AJ, Ziegler JT, Taylor KD, Haffner SM, et al. Genome-wide association study and follow-up analysis of adiposity traits in Hispanic Americans: the IRAS Family Study. Obesity (Silver Spring). 2009;17:1932–41. 140. Fox CS, Liu Y, White CC, Feitosa M, Smith A V, Heard-Costa N, Lohman K, Consortium G, Consortium M, Consortium G, et al. Genome-wide association for abdominal subcutaneous and visceral adipose reveals a novel locus for visceral fat in women. PLoS Genet. 2012;8:e1002695. 141. Plourde M, Vohl MC, Bellis C, Carless M, Dyer T, Dolley G, Marette A, Despres JP, Bouchard C, Blangero J, et al. A variant in the LRRFIP1 gene is associated with adiposity and inflammation. Obesity (Silver Spring). 2013;21:185– 92. 142. Tan LJ, Zhu H, He H, Wu KH, Li J, Chen XD, Zhang JG, Shen H, Tian Q, Krousel-Wood M, et al. Replication of 6 obesity genes in a meta-analysis of genome-wide association studies from diverse ancestries. PLoS One. 2014;9:e96149. 143. Sung YJ, Pérusse L, Sarzynski MA, Fornage M, Sidney S, Sternfeld B, Rice T, Terry JG, Jacobs DR, Katzmarzyk P, et al. Genome-wide association studies suggest sex-specific loci associated with abdominal and visceral fat. Int J Obes (Lond). 2016;40:662–74. 144. Horne BD, Anderson JL, Carlquist JF, Muhlestein JB, Renlund DG, Bair TL, Pearson RR, Camp NJ. Generating genetic risk scores intermediate phenotypes for use in association studies of clinically significant endpoints. Ann Hum Genet. 2005;69:176–86. 145. Belsky DW, Moffitt TE, Sugden K, Williams B, Houts R, McCarthy J, Caspi A. Development and evaluation of a genetic risk score for obesity. Biodemography Soc Biol. 2013;59:85–100. 146. Li S, Zhao JH, Luan J, Luben RN, Rodwell SA, Khaw KT, Ong KK, Wareham NJ, Loos RJF. Cumulative effects and predictive value of common obesity-susceptibility variants identified by genome-wide association studies. Am J Clin Nutr. 2010;91:184–90. 147. Ahmad S, Zhao W, Renström F, Rasheed A, Samuel M, Zaidi M, Shah N, Mallick NH, Zaman KS, Ishaq M, et al. Physical activity, smoking, and genetic predisposition to obesity in people from Pakistan: the PROMIS study. BMC Med Genet. 2015;16:114–25. 148. Renström F, Payne F, Nordström A, Brito EC, Rolandsson O, Hallmans G, Barroso I, Nordström P, Franks PW. Replication and extension of genome-wide association study results for obesity in 4923 adults from Northern Sweden. Hum Mol Genet. 2009;18:1489–96. 149. Klimentidis YC, Chen GB, López-Alarcón M, Harris JJ, Duarte CW, Fernández JR. Associations of obesity genes with obesity-related outcomes in multiethnic children. Arch Med Res. 2011;42:509–14. 150. Peterson RE, Maes HH, Holmans P, Sanders AR, Levinson DF, Shi J, Kendler KS, Gejman P V., Webb BT. Genetic risk sum score comprised of common polygenic variation is associated with body mass index. Hum Genet. 2011;129:221–30. 151. Lemas DJ, Klimentidis YC, Wiener HH, O’Brien DM, Hopkins SE, Allison DB, Fernandez JR, Tiwari HK, Boyer BB. Obesity polymorphisms identified in genome-wide association studies interact with n-3 polyunsaturated fatty acid intake and modify the genetic association with adiposity phenotypes in Yup’ik people. Genes Nutr. 2013;8:495–505. 152. Rukh G, Sonestedt E, Melander O, Hedblad B, Wirfält E, Ericson U, Orho-Melander M. Genetic susceptibility to obesity and diet intakes: association and interaction analyses in the Malmö Diet and Cancer Study. Genes Nutr. 2013;8:535–47. 153. Nettleton JA, Follis JL, Ngwa JS, Smith CE, Ahmad S, Tanaka T, Wojczynski MK, Voortman T, Lemaitre RN, Kristiansson K, et al. Gene x dietary pattern interactions in obesity: analysis of up to 68317 adults of European

363 References ancestry. Hum Mol Genet. 2015;24:4728–38. 154. Qi Q, Li Y, Chomistek AK, Kang JH, Curhan GC, Pasquale LR, Willett WC, Rimm EB, Hu FB, Qi L. Television watching, leisure time physical activity, and the genetic predisposition in relation to body mass index in women and men. Circulation. 2012;126:1821–7. 155. Qi Q, Chu AY, Kang JH, Jensen MK, Curhan GC, Pasquale LR, Ridker PM, Hunter DJ, Willett WC, Rimm EB, et al. Sugar-sweetened beverages and genetic risk of obesity. N Engl J Med. 2012;367:1387–96. 156. Ahmad S, Rukh G, Varga T V., Ali A, Kurbasic A, Shungin D, Ericson U, Koivula RW, Chu AY, Rose LM, et al. Gene x physical activity interactions in obesity: combined analysis of 111,421 individuals of European ancestry. PLoS Genet. 2013;9:e1003607. 157. Jääskeläinen T, Paananen J, Lindström J, Eriksson JG, Tuomilehto J, Uusitupa M. Genetic predisposition to obesity and lifestyle factors-the combined analyses of twenty-six known BMI- and fourteen known waist:hip ratio (WHR)-associated variants in the Finnish Diabetes Prevention Study. Br J Nutr. 2013;110:1856–65. 158. Mägi R, Manning S, Yousseif A, Pucci A, Santini F, Karra E, Querci G, Pelosini C, McCarthy MI, Lindgren CM, et al. Contribution of 32 GWAS-identified common variants to severe obesity in European adults referred for bariatric surgery. PLoS One. 2013;8:e70735. 159. Martínez-García F, Mansego ML, Rojo-Martínez G, De Marco-Solar G, Morcillo S, Soriguer F, Redón J, Pineda Alonso M, Martín-Escudero JC, Cooper RS, et al. Impact of obesity-related genes in Spanish population. BMC Genet. 2013;14:111. 160. Goumidi L, Cottel D, Dallongeville J, Amouyel P, Meirhaeghe A. Effects of established BMI-associated loci on obesity-related traits in a French representative population sample. BMC Genet. 2014;15:62. 161. Qi Q, Chu AY, Kang JH, Huang J, Rose LM, Jensen MK, Liang L, Curhan GC, Pasquale LR, Wiggs JL, et al. Fried food consumption, genetic risk, and body mass index: gene-diet interaction analysis in three US cohort studies. BMJ. 2014;19:g1610. 162. Domingue BW, Belsky DW, Harris KM, Smolen A, McQueen MB, Boardman JD. Polygenic risk predicts obesity in both white and black young adults. PLoS One. 2014;9:5–12. 163. Fox CS, Pencina MJ, Heard-Costa NL, Shrader P, Jaquish C, O’Donnell CJ, Vasan RS, Cupples LA, D’Agostino RB. Trends in the association of parental history of obesity over 60 years. Obesity. 2014;22:919–24. 164. Sandholt CH, Allin KH, Toft U, Borglykke A, Ribel-Madsen R, Sparso T, Justesen JM, Harder MN, Jørgensen T, Hansen T, et al. The effect of GWAS identified BMI loci on changes in body weight among middle-aged Danes during a five-year period. Obesity. 2014;22:901–8. 165. Hung CF, Breen G, Czamara D, Corre T, Wolf C, Kloiber S, Bergmann S, Craddock N, Gill M, Holsboer F, et al. A genetic risk score combining 32 SNPs is associated with body mass index and improves obesity prediction in people with major depressive disorder. BMC Med. 2015;13:86. 166. Cheung CYY, Tso AWK, Cheung BMY, Xu A, Ong KL, Fong CHY, Wat NMS, Janus ED, Sham PC, Lam KSL. Obesity susceptibility genetic variants identified from recent genome-wide association studies: Implications in a Chinese population. J Clin Endocrinol Metab. 2010;95:1395–403. 167. Manolio TA, Collins FS, Cox NJ, Goldstein DB, Hindorff LA, Hunter DJ, McCarthy MI, Ramos EM, Cardon LR, Chakravarti A, et al. Finding the missing heritability of complex diseases. Nature. 2009;461:747–53. 168. Heianza Y, Qi L. Gene-diet interaction and precision nutrition in obesity. Int J Mol Sci. 2017;18:787–802. 169. Simopoulos AP. Nutrigenetics/Nutrigenomics. Annu Rev Public Health. 2010;31:53–68. 170. Vliet-Ostaptchouk J V., Snieder H, Lagou V. Gene–lifestyle interactions in obesity. Curr Nutr Rep. 2012;1:184– 96.

364 References

171. Corella D, Arnett DK, Tucker KL, Kabagambe EK, Tsai M, Parnell LD, Lai C-Q, Lee Y-C, Warodomwichit D, Hopkins PN, et al. A high intake of saturated fatty acids strengthens the association between the fat mass and obesity-associated gene and BMI. J Nutr. 2011;141:2219–25. 172. Qi Q, Kilpeläinen TO, Downer MK, Tanaka T, Smith CE, Sluijs I, Sonestedt E, Chu AY, Renström F, Lin X, et al. FTO genetic variants, dietary intake and body mass index: insights from 177,330 individuals. Hum Mol Genet. 2014;23:6961–72. 173. Corella D, Tai ES, Sorlí J V, Chew SK, Coltell O, Sotos-Prieto M, García-Rios A, Estruch R, Ordovas JM. Association between the APOA2 promoter polymorphism and body weight in Mediterranean and Asian populations: replication of a gene-saturated fat interaction. Int J Obes. 2011;35:666–75. 174. Doo M, Won S, Kim Y. Association between the APOB rs1469513 polymorphism and obesity is modified by dietary fat intake in Koreans. Nutrition. 2015;31:653–8. 175. Ma Y, Tucker KL, Smith CE, Lee YC, Huang T, Richardson K, Parnell LD, Lai CQ, Young KL, Justice AE, et al. Lipoprotein lipase variants interact with polyunsaturated fatty acids for obesity traits in women: replication in two populations. Nutr Metab Cardiovasc Dis. 2014;24:1323–9. 176. Vaughan LK, Wiener HW, Aslibekyan S, Allison DB, Havel PJ, Stanhope KL, O’Brien DM, Hopkins SE, Lemas DJ, Boyer BB, et al. Linkage and association analysis of obesity traits reveals novel loci and interactions with dietary n-3 fatty acids in an Alaska Native (Yup’ik) population. Metabolism. 2015;64:689–97. 177. Junyent M, Parnell LD, Lai CQ, Arnett DK, Tsai MY, Kabagambe EK, Straka RJ, Province M, An P, Smith CE, et al. ADAM17_i33708A>G polymorphism interacts with dietary n-6 polyunsaturated fatty acids to modulate obesity risk in the Genetics of Lipid Lowering Drugs and Diet Network study. Nutr Metab Cardiovasc Dis. 2010;20:698–705. 178. Smith CE, Tucker KL, Yiannakouris N, Garcia-Bailo B, Mattei J, Lai CQ, Parnell LD, Ordovas JM. Perilipin polymorphism interacts with dietary carbohydrates to modulate anthropometric traits in hispanics of Caribbean origin. J Nutr. 2008;138:1852–8. 179. Patel CJ, Rehkopf DH, Leppert JT, Bortz WM, Cullen MR, Chertow GM, Ioannidis JPA. Systematic evaluation of environmental and behavioural factors associated with all-cause mortality in the United States national health and nutrition examination survey. Int J Epidemiol. 2013;42:1795–810. 180. Petkeviciene J, Smalinskiene A, Klumbiene J, Petkevicius V, Kriaucioniene V, Lesauskaite V. Physical activity, but not dietary intake, attenuates the effect of the FTO rs9939609 polymorphism on obesity and metabolic syndrome in Lithuanian adult population. Public Health. 2016;135:23–9. 181. Celis-Morales C, Marsaux CFM, Livingstone KM, Navas-Carretero S, San-Cristobal R, O’Donovan CB, Forster H, Woolhead C, Fallaize R, Macready AL, et al. Physical activity attenuates the effect of the FTO genotype on obesity traits in European adults: The Food4Me study. Obesity. 2016;24:962–9. 182. Jonsson A, Renström F, Lyssenko V, Brito EC, Isomaa B, Berglund G, Nilsson PM, Groop L, Franks PW. Assessing the effect of interaction between an FTO variant (rs9939609) and physical activity on obesity in 15,925 Swedish and 2,511 Finnish adults. Diabetologia. 2009;52:1334–8. 183. Lappalainen TJ, Tolppanen A-M, Kolehmainen M, Schwab U, Lindström J, Tuomilehto J, Pulkkinen L, Eriksson JG, Laakso M, Gylling H, et al. The common variant in the FTO gene did not modify the effect of lifestyle changes on body weight: the Finnish Diabetes Prevention Study. Obesity (Silver Spring). 2009;17:832–6. 184. Graff M, Scott RA, Justice AE, Young KL, Feitosa F, Barata L, Winkler TW, Chu AY, Mahajan A, Hadley D, et al. Genome-wide physical activity interactions in adiposity - A meta-analysis of 200,452 adults. PLoS Genet. 2017;13:e1006528. 185. Klimentidis Y, Arora A, Chougule A, Zhou J, Raichlen D. FTO association and interaction with time spent sitting.

365 References

Int J Obes. 2016;40:411–6. 186. Ahmad S, Zhao W, Renström F, Rasheed A, Zaidi M, Samuel M, Shah N, Mallick NH, Shungin D, Zaman KS, et al. A novel interaction between the FLJ33534 locus and smoking in obesity: a genome-wide study of 14,131 Pakistani adults. Int J Obes. 2016;40:186–90. 187. Edwards TL, Velez Edwards DR, Villegas R, Cohen SS, Buchowski MS, Fowke JH, Schlundt D, Long JR, Cai Q, Zheng W, et al. HTR1B, ADIPOR1, PPARGC1A, and CYP19A1 and obesity in a Cohort of Caucasians and African Americans: An evaluation of gene-environment interactions and candidate genes. Am J Epidemiol. 2012;175:11–21. 188. Dashti HS, Follis JL, Smith CE, Tanaka T, Garaulet M, Gottlieb DJ, Hruby A, Jacques PF, Kiefte-De Jong JC, Lamon-Fava S, et al. Gene-environment interactions of circadian-related genes for cardiometabolic traits. Diabetes Care. 2015;38:1456–66. 189. Corella D, Carrasco P, Sorlí J V., Coltell O, Ortega-Azorín C, Guillén M, González JI, Sáiz C, Estruch R, Ordovas JM. Education modulates the association of the FTO rs9939609 polymorphism with body mass index and obesity risk in the Mediterranean population. Nutr Metab Cardiovasc Dis. 2012;22:651–8. 190. Lagou V, Liu G, Zhu H, Stallmann-Jorgensen IS, Gutin B, Dong Y, Snieder H. Lifestyle and socioeconomic-status modify the effects of ADRB2 and NOS3 on adiposity in European-American and African-American adolescents. Obesity. 2011;19:595–603. 191. Waken R, de las Fuentes L, Rao DC. A Review of the genetics of hypertension with a focus on gene- environment interactions. Curr Hypertens Rep. 2017;19:23–31. 192. Kushner RF. Weight loss strategies for treatment of obesity. Prog Cardiovasc Dis. 2014;56:465–72. 193. Abete I, Astrup A, Martinez JA, Thorsdottir I, Zulet MA. Obesity and the metabolic syndrome: role of different dietary macronutrient distribution patterns and specific nutritional components on weight loss and maintenance. Nutr Rev. 2010;68:214–31. 194. Perez-Cornago A, Ramirez MJ, Zulet MA, Martinez JA. Effect of dietary restriction on peripheral monoamines and anxiety symptoms in obese subjects with metabolic syndrome. Psychoneuroendocrinology. 2014;47:98–106. 195. Burmeister JM, Carels RA. Television use and binge eating in adults seeking weight loss treatment. Eat Behav. 2014;15:83–6. 196. Mariman EC. Human biology of weight maintenance after weight loss. J Nutrigenet Nutrigenomics. 2012;5:13– 25. 197. Goni L, Cuervo M, Milagro FI, Mart JA. Future perspectives of personalized weight loss interventions based on nutrigenetic, epigenetic, and metagenomic data. J Nutr. 2016;146:905–12. 198. Matarese LE, Pories WJ. Adult weight loss diets: metabolic effects and outcomes. Nutr Clin Pract. 2014;29:759–67. 199. Lee JS, Kawakubo K, Inoue S, Akabayashi A. Effect of beta(3)-adrenergic receptor gene polymorphism on body weight change in middle-aged, overweight women. Environ Health Prev Med. 2006;11:69–74. 200. Yamakita M, Ando D, Tang S, Yamagata Z. The Trp64Arg polymorphism of the beta3-adrenergic receptor gene is associated with weight changes in obese Japanese men: a 4-year follow-up study. J Physiol Anthropol. 2010;29:133–9. 201. Tchernof A, Starling RD, Turner A, Shuldiner AR, Walston JD, Silver K, Poehlman ET. Impaired capacity to lose visceral adipose tissue during weight reduction in obese postmenopausal women with the Trp64Arg beta3- adrenoceptor gene variant. Diabetes. 2000;49:1709–13. 202. Sakane N, Yoshida T, Umekawa T, Kogure A, Kondo M. Beta2-adrenoceptor gene polymorphism and obesity. Lancet. 1999;353:1976.

366 References

203. Szendrei B, González-Lamuño D, Amigo T, Wang G, Pitsiladis Y, Benito PJ, Gomez-Candela C, Calderón FJ, Cupeiro R, PRONAF Study Group. Influence of ADRB2 Gln27Glu and ADRB3 Trp64Arg polymorphisms on body weight and body composition changes after a controlled weight-loss intervention. Appl Physiol Nutr Metab. 2016;41:307–14. 204. Cha MH, Shin HD, Kim KS, Lee BH, Yoon Y. The effects of uncoupling protein 3 haplotypes on obesity phenotypes and very low-energy diet-induced changes among overweight Korean female subjects. Metabolism. 2006;55:578–86. 205. Yoon Y, Park BL, Cha MH, Kim KS, Cheong HS, Choi YH, Shin HD. Effects of genetic polymorphisms of UCP2 and UCP3 on very low calorie diet-induced body fat reduction in Korean female subjects. Biochem Biophys Res Commun. 2007;359:451–6. 206. Nagai N, Sakane N, Kotani K, Hamada T, Tsuzaki K, Moritani T. Uncoupling protein 1 gene -3826 A/G polymorphism is associated with weight loss on a short-term, controlled-energy diet in young women. Nutr Res. 2011;31:255–61. 207. De Luis DA, Aller R, Izaola O, De La Fuente B, Conde R, Eiros Bouza JM. Relation of -55CT polymorphism of UCP3 gene with weight loss and metabolic changes after a high monounsaturated fat diet in obese non diabetic patients. Eur Rev Med Pharmacol Sci. 2013;17:2810–5. 208. Argiles JM, Busquets S, Lopez-Soriano FJ. The role of uncoupling proteins in pathophysiological states. Biochem Biophys Res Commun. 2002;293:1145–52. 209. Santos MJL. Leptin-melanocortin system, body weight regulation and obesity. Rev Med Chil. 2009;137:1225– 34. 210. Sainz N, Barrenetxe J, Moreno-Aliaga MJ, Martinez JA. Leptin resistance and diet-induced obesity: central and peripheral actions of leptin. Metabolism. 2015;64:35–46. 211. Mammes O, Betoulle D, Aubert R, Giraud V, Tuzet S, Petiet A, Colas-Linhart N, Fumeron F. Novel polymorphisms in the 5’ region of the LEP gene: association with leptin levels and response to low-calorie diet in human obesity. Diabetes. 1998;47:487–9. 212. Zacharova J, Chiasson JL, Laakso M. Leptin receptor gene variation predicts weight change in subjects with impaired glucose tolerance. Obes Res. 2005;13:501–6. 213. de Luis DA, Aller R, Izaola O, Sagrado MG, Conde R. Influence of Lys656Asn polymorphism of leptin receptor gene on leptin response secondary to two hypocaloric diets: a randomized clinical trial. Ann Nutr Metab. 2008;52:209–14. 214. Pan Q, Delahanty LM, Jablonski KA, Knowler WC, Kahn SE, Florez JC, Franks PW, Group DPPR. Variation at the melanocortin 4 receptor gene and response to weight-loss interventions in the Diabetes Prevention program. Obesity (Silver Spring). 2013;21:E520-6. 215. Jaaskelainen T, Paananen J, Lindstrom J, Eriksson JG, Tuomilehto J, Uusitupa M, Group FDPS. Genetic predisposition to obesity and lifestyle factors-the combined analyses of twenty-six known BMI- and fourteen known waist:hip ratio (WHR)-associated variants in the Finnish Diabetes Prevention Study. Br J Nutr. 2013;110:1856–65. 216. Nelson DL, Gehlert DR. Central nervous system biogenic amine targets for control of appetite and energy expenditure. Endocrine. 2006;29:49–60. 217. Cameron JD, Riou ME, Tesson F, Goldfield GS, Rabasa-Lhoret R, Brochu M, Doucet E. The TaqIA RFLP is associated with attenuated intervention-induced body weight loss and increased carbohydrate intake in post- menopausal obese women. Appetite. 2013;60:111–6. 218. Pooley EC, Fairburn CG, Cooper Z, Sodhi MS, Cowen PJ, Harrison PJ. A 5-HT2C receptor promoter

367 References polymorphism (HTR2C - 759C/T) is associated with obesity in women, and with resistance to weight loss in heterozygotes. Am J Med Genet. 2004;126B:124–7. 219. Watkins BA, Kim J. The endocannabinoid system: directing eating behavior and macronutrient metabolism. Front Psychol. 2015;5:1506. 220. de Luis D, Aller R, Izaola O, Conde R, de la Fuente B, Gonzalez Sagrado M. Genetic variation in the endocannabinoid degrading enzyme fatty acid amide hydrolase (FAAH) and their influence on weight loss and insulin resistance under a high monounsaturated fat hypocaloric diet. J Diabetes Complications. 2013;27:235–9. 221. Ketterer C, Heni M, Stingl K, Tschritter O, Linder K, Wagner R, Machicao F, Haring HU, Preissl H, Staiger H, et al. Polymorphism rs3123554 in CNR2 reveals gender-specific effects on body weight and affects loss of body weight and cerebral insulin action. Obesity (Silver Spring). 2014;22:925–31. 222. Delahanty LM, Pan Q, Jablonski KA, Watson KE, McCaffery JM, Shuldiner A, Kahn SE, Knowler WC, Florez JC, Franks PW, et al. Genetic predictors of weight loss and weight regain after intensive lifestyle modification, metformin treatment, or standard care in the Diabetes Prevention Program. Diabetes Care. 2012;35:363–6. 223. Basile KJ, Johnson ME, Xia Q, Grant SF. Genetic susceptibility to type 2 diabetes and obesity: follow-up of findings from genome-wide association studies. Int J Endocrinol. 2014;2014:769671. 224. Grau K, Cauchi S, Holst C, Astrup A, Martinez JA, Saris WH, Blaak EE, Oppert JM, Arner P, Rossner S, et al. TCF7L2 rs7903146-macronutrient interaction in obese individuals’ responses to a 10-wk randomized hypoenergetic diet. Am J Clin Nutr. 2010;91:472–9. 225. Haupt A, Thamer C, Heni M, Ketterer C, Machann J, Schick F, Machicao F, Stefan N, Claussen CD, Haring HU, et al. Gene variants of TCF7L2 influence weight loss and body composition during lifestyle intervention in a population at risk for type 2 diabetes. Diabetes. 2010;59:747–50. 226. Mattei J, Qi Q, Hu FB, Sacks FM, Qi L. TCF7L2 genetic variants modulate the effect of dietary fat intake on changes in body composition during a weight-loss intervention. Am J Clin Nutr. 2012;96:1129–36. 227. Heni M, Herzberg-Schafer S, Machicao F, Haring HU, Fritsche A. Dietary fiber intake modulates the association between variants in TCF7L2 and weight loss during a lifestyle intervention. Diabetes Care. 2012;35:e24-2012. 228. Wei H, Averill MM, McMillen TS, Dastvan F, Mitra P, Subramanian S, Tang C, Chait A, LeBoeuf RC. Modulation of adipose tissue lipolysis and body weight by high-density lipoproteins in mice. Nutr Diabetes. 2014;4:e108. 229. Aberle J, Evans D, Beil FU, Seedorf U. A polymorphism in the apolipoprotein A5 gene is associated with weight loss after short-term diet. Clin Genet. 2005;68:152–4. 230. Adamo KB, Dent R, Langefeld CD, Cox M, Williams K, Carrick KM, Stuart JS, Sundseth SS, Harper ME, McPherson R, et al. Peroxisome proliferator-activated receptor gamma 2 and acyl-CoA synthetase 5 polymorphisms influence diet response. Obesity (Silver Spring). 2007;15:1068–75. 231. Ruiz JR, Larrarte E, Margareto J, Ares R, Alkorta P, Labayen I. Preliminary findings on the role of PLIN1 polymorphisms on body composition and energy metabolism response to energy restriction in obese women. Br J Nutr. 2011;106:486–90. 232. Corella D, Qi L, Sorli J V, Godoy D, Portoles O, Coltell O, Greenberg AS, Ordovas JM. Obese subjects carrying the 11482G>A polymorphism at the perilipin locus are resistant to weight loss after dietary energy restriction. J Clin Endocrinol Metab. 2005;90:5121–6. 233. Jang Y, Kim OY, Lee JH, Koh SJ, Chae JS, Kim JY, Park S, Cho H, Lee JE, Ordovas JM. Genetic variation at the perilipin locus is associated with changes in serum free fatty acids and abdominal fat following mild weight loss. Int J Obes (Lond). 2006;30:1601–8. 234. Soenen S, Mariman EC, Vogels N, Bouwman FG, den Hoed M, Brown L, Westerterp-Plantenga MS.

368 References

Relationship between perilipin gene polymorphisms and body weight and body composition during weight loss and weight maintenance. Physiol Behav. 2009;96:723–8. 235. de Luis D, Aller R, Izaola O, Sagrado MG, de la Fuente B, Conde R, Primo D. Effect of fatty acid-binding protein 2 Ala54Thr genotype on weight loss and cardiovascular risk factors after a high-polyunsaturated fat diet in obese patients. J Investig Med. 2012;60:1194–8. 236. Stocks T, Angquist L, Banasik K, Harder MN, Taylor MA, Hager J, Arner P, Oppert JM, Martinez JA, Polak J, et al. TFAP2B influences the effect of dietary fat on weight loss under energy restriction. PLoS One. 2012;7:e43212. 237. Zhang X, Qi Q, Zhang C, Smith SR, Hu FB, Sacks FM, Bray GA, Qi L. FTO genotype and 2-year change in body composition and fat distribution in response to weight-loss diets: the POUNDS LOST Trial. Diabetes. 2012;61:3005– 11. 238. Mitchell JA, Church TS, Rankinen T, Earnest CP, Sui X, Blair SN. FTO genotype and the weight loss benefits of moderate intensity exercise. Obesity (Silver Spring). 2010;18:641–3. 239. de Luis DA, Aller R, Conde R, Izaola O, Gonzalez Sagrado M, Castrodeza Sanz J. The rs9939609 gene variant in FTO modified the metabolic response of weight loss after a 3-month intervention with a hypocaloric diet. J Investig Med. 2013;61:22–6. 240. Livingstone KM, Celis-Morales C, Papandonatos GD, Erar B, Florez JC, Jablonski KA, Razquin C, Marti A, Heianza Y, Huang T, et al. FTO genotype and weight loss: systematic review and meta-analysis of 9563 individual participant data from eight randomised controlled trials. BMJ. 2016;354:i4707. 241. Gomez-Abellan P, Madrid JA, Ordovas JM, Garaulet M. Chronobiological aspects of obesity and metabolic syndrome. Endocrinol Nutr. 2012;59:50–61. 242. Garaulet M, Sanchez-Moreno C, Smith CE, Lee YC, Nicolas F, Ordovas JM. Ghrelin, sleep reduction and evening preference: relationships to CLOCK 3111 T/C SNP and weight loss. PLoS One. 2011;6:e17435. 243. Sacks FM, Bray GA, Carey VJ, Smith SR, Ryan DH, Anton SD, McManus C, Champagne CM, Bishop LM, Laranjo N, et al. Comparison of weight-loss diets with different compositions of fat, protein, and carbohydrates. N Engl J Med. 2009;360:859–73. 244. Qi Q, Bray GA, Smith SR, Hu FB, Sacks FM, Qi L. Insulin receptor substrate 1 gene variation modifies insulin resistance response to weight-loss diets in a 2-year randomized trial: the Preventing Overweight Using Novel Dietary Strategies (POUNDS LOST) trial. Circulation. 2011;124:563–71. 245. Heianza Y, Ma W, Huang T, Wang T, Zheng Y, Smith SR, Bray GA, Sacks FM, Qi L. Macronutrient intake- associated FGF21 genotype modifies effects of weight-loss diets on 2-year changes of central adiposity and body composition: The POUNDS lost trial. Diabetes Care. 2016;39:1909–14. 246. Petersen M, Taylor MA, Saris WHM, Verdich C, Toubro S, Macdonald I, Rössner S, Stich V, Guy-Grand B, Langin D, et al. Randomized, multi-center trial of two hypo-energetic diets in obese subjects: high- versus low-fat content. Int J Obes. 2005;30:552–60. 247. Papadaki A, Linardakis M, Plada M, Larsen TM, Damsgaard CT, van Baak MA, Jebb S, Pfeiffer AFH, Martinez JA, Handjieva-Darlenska T, et al. Impact of weight loss and maintenance with ad libitum diets varying in protein and glycemic index content on metabolic syndrome. Nutrition. 2014;30:410–7. 248. Larsen LH, Ängquist L, Vimaleswaran KS, Hager J, Viguerie N, Loos RJF, Handjieva-Darlenska T, Jebb SA, Kunešova M, Larsen TM, et al. Analyses of single nucleotide polymorphisms in selected nutrientsensitive genes in weight-regain prevention: The DIOGENES study. Am J Clin Nutr. 2012;95:1254–60. 249. Grau K, Hansen T, Holst C, Astrup A, Saris WH, Arner P, Rossner S, Macdonald I, Polak J, Oppert JM, et al. Macronutrient-specific effect of FTO rs9939609 in response to a 10-week randomized hypo-energetic diet among

369 References obese Europeans. Int J Obes (Lond). 2009;33:1227–34. 250. Garaulet M, Smith CE, Hernandez-Gonzalez T, Lee YC, Ordovas JM. PPARgamma Pro12Ala interacts with fat intake for obesity and weight loss in a behavioural treatment based on the Mediterranean diet. Mol Nutr Food Res. 2011;55:1771–9. 251. Qi Q, Bray GA, Hu FB, Sacks FM, Qi L. Weight-loss diets modify glucose-dependent insulinotropic polypeptide receptor rs2287019 genotype effects on changes in body weight, fasting glucose, and insulin resistance: the Preventing Overweight Using Novel Dietary Strategies trial. Am J Clin Nutr. 2012;95:506–13. 252. Xu M, Qi Q, Liang J, Bray GA, Hu FB, Sacks FM, Qi L. Genetic determinant for amino acid metabolites and changes in body weight and insulin resistance in response to weight-loss diets: The preventing overweight using novel dietary strategies (POUNDS LOST) trial. Circulation. 2013;127:1283–9. 253. Huang T, Qi Q, Li Y, Hu FB, Bray GA, Sacks FM, Williamson DA, Qi L. FTO genotype, dietary protein, and change in appetite: the Preventing Overweight Using Novel Dietary Strategies trial. Am J Clin Nutr. 2014;99:1126–30. 254. Mirzaei K, Xu M, Qi Q, de Jonge L, Bray GA, Sacks F, Qi L. Variants in glucose- and circadian rhythm-related genes affect the response of energy expenditure to weight-loss diets: the POUNDS LOST Trial. Am J Clin Nutr. 2014;99:392–9. 255. Lin X, Qi Q, Zheng Y, Huang T, Lathrop M, Zelenika D, Bray GA, Sacks FM, Liang L, Qi L. Neuropeptide Y genotype, central obesity, and abdominal fat distribution: The POUNDS LOST trial. Am J Clin Nutr. 2015;102:514–9. 256. Larsen SC, Ängquist L, Østergaard JN, Ahluwalia TS, Vimaleswaran KS, Roswall N, Mortensen LM, Nielsen BM, Tjønneland A, Wareham NJ, et al. Intake of total and subgroups of fat minimally affect the associations between selected single nucleotide polymorphisms in the PPARγ pathway and changes in anthropometry among European adults from cohorts of the DiOGenes study. J Nutr. 2016;146:603–11. 257. Loria-Kohen V, Espinosa-Salinas I, Marcos-Pasero H, Lourenço-Nogueira T, Herranz J, Molina S, Reglero G, Ramirez de Molina A. Polymorphism in the CLOCK gene may influence the effect of fat intake reduction on weight loss. Nutrition. 2016;32:453–60. 258. Hruby A, Hu FB. The epidemiology of obesity: a big picture. Pharmacoeconomics. 2015;33:673–89. 259. Feingold KR, Grunfeld C. Introduction to lipids and lipoproteins. Endotext. 2015. 260. Fredrickson DS, Lees RS. A system for phenotyping hyperlipoproteinemia. Circulation. 1965;31:321–7. 261. Patni N, Ahmad Z, Wilson DP. Genetics and dyslipidemia. Endotext. 2000. 262. Mataix Verdú J, Ramírez JA, Solá R. Enfermedades del sistema circulatorio. In: Mataix Verdú J, editor. Nutrición y Alimentación Humana. 2nd ed. Madrid; 2015. p. 1469–517. 263. Catapano AL, Graham I, De Backer G, Wiklund O, Chapman MJ, Drexel H, Hoes AW, Jennings CS, Landmesser U, Pedersen TR, et al. 2016 ESC/EAS guidelines for the management of dyslipidaemias. Eur Heart J. 2016;37:2999– 3058l. 264. Stone NJ, Ronbinson J, Lichtenstein AH, Merz CNB, Blum CB, Eckel RH, Goldberg AC, Gordon D, Levy D, Lloyd- Jones DM, et al. 2013 ACC/AHA guideline on the tratment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/ American Heart Association task force on practice guidelines. Circulation. 2013;67:246–7. 265. Jellinger PS, Handelsman Y, Rosenblit PD, Bloomgarden ZT, Fonseca VA, Garber AJ, Grunberger G, Guerin CK, Bell DSH, Mechanick JI, et al. American Association of Clinical Endocrinologists and American College of Endocrinology guidelines for management of dyslipidemia and prevention of cardiovascular disease. Endocr Pract. 2017;23:1–87. 266. Mora S, Kamstrup PR, Rifai N, Nordestgaard BG, Buring JE, Ridker PM. Lipoprotein(a) and risk of type 2

370 References diabetes. Clin Chem. 2010;56:1252–60. 267. Greenhill C. Cancer: HDL cholesterol and cancer risk. Nat Rev Gastroenterol Hepatol. 2011;8:299–299. 268. Touvier M, Fassier P, His M, Norat T, Chan DSM, Blacher J, Hercberg S, Galan P, Druesne-Pecollo N, Latino- Martel P. Cholesterol and breast cancer risk: a systematic review and meta-analysis of prospective studies. Br J Nutr. 2015;114:347–57. 269. World Health Organization (WHO). Raised cholesterol. Situation and trends [Internet]. [cited 2017 Jan 1]. Available from: http://www.who.int/gho/ncd/risk_factors/cholesterol_text/en/ 270. Farzadfar F, Finucane MM, Danaei G, Pelizzari PM, Cowan MJ, Paciorek CJ, Singh GM, Lin JK, Stevens GA, Riley LM, et al. National, regional, and global trends in serum total cholesterol since 1980: systematic analysis of health examination surveys and epidemiological studies with 321 country-years and 3·0 million participants. Lancet. 2011;377:578–86. 271. Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K, Adair-Rohani H, Amann M, Anderson HR, Andrews KG, Aryee M, et al. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380:2224–60. 272. Sharma K, Baliga RR. Genetics of dyslipidemia and ischemic heart disease. Curr Cardiol Rep. 2017;19. 273. Watts GF, Gidding S, Wierzbicki AS, Toth PP, Alonso R, Brown WV, Bruckert E, Defesche J, Lin KK, Livingston M, et al. Integrated guidance on the care of familial hypercholesterolaemia from the International FH Foundation. Eur J Prev Cardiol. 2015;22:849–54. 274. Nordestgaard BG, Chapman MJ, Humphries SE, Ginsberg HN, Masana L, Descamps OS, Wiklund O, Hegele RA, Raal FJ, Defesche JC, et al. Familial hypercholesterolaemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease. Eur Heart J. 2013;34:3478–90. 275. Naukkarinen J, Nilsson E, Koistinen HA, Söderlund S, Lyssenko V, Vaag A, Poulsen P, Groop L, Taskinen MR, Peltonen L. Functional variant disrupts insulin induction of USF1: mechanism for USF1-associated dyslipidemias. Circ Cardiovasc Genet. 2009;2:522–9. 276. Iatan I, Choi HY, Ruel I, Linga Reddy MVP, Kil H, Lee J, Abu Odeh M, Salah Z, Abu-Remaileh M, Weissglas- Volkov D, et al. The WWOX gene modulates high-density lipoprotein and lipid metabolism. Circ Cardiovasc Genet. 2014;7:491–504. 277. Lin S, Dai R, Lin R. A meta-analytic evaluation of cholesteryl ester transfer protein (CETP) C-629A polymorphism in association with coronary heart disease risk and lipid changes. Oncotarget. 2016;8:2153–63. 278. Sacks FM. The crucial roles of apolipoproteins E and C-III in apoB lipoprotein metabolism in normolipidemia and hypertriglyceridemia. Curr Opin Lipidol. 2015;26:56–63. 279. Helgadottir A, Gretarsdottir S, Thorleifsson G, Hjartarson E, Sigurdsson A, Magnusdottir A, Jonasdottir A, Kristjansson H, Sulem P, Oddsson A, et al. Variants with large effects on blood lipids and the role of cholesterol and triglycerides in coronary disease. Nat Genet. 2016;48:634–9. 280. Hegele RA. Plasma lipoproteins: genetic influences and clinical implications. Nat Rev Genet. 2009;10:109–21. 281. Musunuru K, Lettre G, Young T, Farlow DN, Pirruccello JP, Ejebe KG, Keating BJ, Yang Q, Chen MH, Lapchyk N, et al. Candidate Gene Association Resource (CARe): design, methods, and proof of concept. Circ Cardiovasc Genet. 2010;3:267–75. 282. Lange LA, Willer CJ, Rich SS. Recent developments in genome and exome-wide analyses of plasma lipids. Curr Opin Lipidol. 2015;26:96–102. 283. Keenan TE, Rader DJ. Genetics of lipid traits and relationship to coronary artery disease. Curr Cardiol Rep.

371 References

2013;15. 284. Saxena R, Voight BF, Lyssenko V, Burtt NP, de Bakker PIW, Chen H, Roix JJ, Kathiresan S, Hirschhorn JN, Daly MJ, et al. Genome-wide association analysis identifies loci for type 2 diabetes and triglyceride levels. Science. 2007;316:1331–6. 285. Surakka I, Horikoshi M, Mägi R, Sarin A-P, Mahajan A, Lagou V, Marullo L, Ferreira T, Miraglio B, Timonen S, et al. The impact of low-frequency and rare variants on lipid levels. Nat Genet. 2015;47:589–97. 286. Kathiresan S, Melander O, Guiducci C, Surti A, Burtt NP, Rieder MJ, Cooper GM, Roos C, Voight BF, Havulinna AS, et al. Six new loci associated with blood low-density lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in humans. Nat Genet. 2008;40:189–97. 287. Willer CJ, Sanna S, Jackson AU, Scuteri A, Bonnycastle LL, Clarke R, Heath SC, Timpson NJ, Najjar SS, Stringham HM, et al. Newly identified loci that influence lipid concentrations and risk of coronary artery disease. Nat Genet. 2008;40:161–9. 288. Kooner JS, Chambers JC, Aguilar-Salinas CA, Hinds DA, Hyde CL, Warnes GR, Gómez Pérez FJ, Frazer KA, Elliott P, Scott J, et al. Genome-wide scan identifies variation in MLXIPL associated with plasma triglycerides. Nat Genet. 2008;40:149–51. 289. Kamatani Y, Matsuda K, Okada Y, Kubo M, Hosono N, Daigo Y, Nakamura Y, Kamatani N. Genome-wide association study of hematological and biochemical traits in a Japanese population. Nat Genet. 2010;42:210–5. 290. Teslovich TM, Musunuru K, Smith A V, Edmondson AC, Stylianou IM, Koseki M, Pirruccello JP, Ripatti S, Chasman DI, Willer CJ, et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature. 2010;466:707–13. 291. Kim YJ, Go MJ, Hu C, Hong CB, Kim YK, Lee JY, Hwang J-Y, Oh JH, Kim D-J, Kim NH, et al. Large-scale genome- wide association studies in east Asians identify new genetic loci influencing metabolic traits. Nat Genet. 2011;43:990–5. 292. Coram MA, Duan Q, Hoffmann TJ, Thornton T, Knowles JW, Johnson NA, Ochs-Balcom HM, Donlon TA, Martin LW, Eaton CB, et al. Genome-wide characterization of shared and distinct genetic components that influence blood lipid levels in ethnically diverse human populations. Am J Hum Genet. 2013;92:904–16. 293. Consortium GLG, Willer CJ, Schmidt EM, Sengupta S, Peloso GM, Gustafsson S, Kanoni S, Ganna A, Chen J, Buchkovich ML, et al. Discovery and refinement of loci associated with lipid levels. Nat Genet. 2013;45:1274–83. 294. Ko A, Cantor RM, Weissglas-Volkov D, Nikkola E, Reddy PMVL, Sinsheimer JS, Pasaniuc B, Brown R, Alvarez M, Rodriguez A, et al. Amerindian-specific regions under positive selection harbour new lipid variants in Latinos. Nat Commun. 2014;5:3983–95. 295. van Leeuwen EM, Sabo A, Bis JC, Huffman JE, Manichaikul A, Smith A V, Feitosa MF, Demissie S, Joshi PK, Duan Q, et al. Meta-analysis of 49,549 individuals imputed with the 1000 Genomes Project reveals an exonic damaging variant in ANGPTL4 determining fasting TG levels. J Med Genet. 2016;53:441–9. 296. Richardson K, Nettleton JA, Rotllan N, Tanaka T, Smith CE, Lai CQ, Parnell LD, Lee YC, Lahti J, Lemaitre RN, et al. Gain-of-function lipoprotein lipase variant rs13702 modulates lipid traits through disruption of a MicroRNA-410 seed site. Am J Hum Genet. 2013;92:5–14. 297. Ortega-Azorín C, Sorlí J V., Estruch R, Asensio EM, Coltell O, González JI, Martínez-González MÁ, Ros E, Salas- Salvadó J, Fitó M, et al. Amino acid change in the carbohydrate response element binding protein is associated with lower triglycerides and myocardial infarction incidence depending on level of adherence to the mediterranean diet in the PREDIMED trial. Circ Cardiovasc Genet. 2014;7:49–58. 298. Grammer TB, Hoffmann MM, Scharnagl H, Kleber ME, Silbernagel G, Pilz S, Tomaschitz A, Lerchbaum E,

372 References

Siekmeier R, März W. Smoking, apolipoprotein E genotypes, and mortality (the Ludwigshafen RIsk and Cardiovascular Health study). Eur Heart J. 2013;34:1298–305. 299. Son KY, Son H-Y, Chae J, Hwang J, Jang S, Yun JM, Cho B, Park JH, Kim J-I, Alwan A, et al. Genetic association of APOA5 and APOE with metabolic syndrome and their interaction with health-related behavior in Korean men. Lipids Health Dis. 2015;14:105. 300. Brahe LK, Ängquist L, Larsen LH, Vimaleswaran KS, Hager J, Viguerie N, Loos RJF, Handjieva-Darlenska T, Jebb SA, Hlavaty P, et al. Influence of SNPs in nutrient-sensitive candidate genes and gene–diet interactions on blood lipids: the DiOGenes study. Br J Nutr. 2013;110:790–6. 301. Volcik KA, Nettleton JA, Ballantyne CM, Boerwinkle E. Peroxisome proliferator – activated receptor alpha genetic variation interacts with n-6 and long-chain n-3 fatty acid intake to affect total cholesterol and LDL- cholesterol concentrations in the Atherosclerosis Risk in Communities Study. 2008;87:1926–31. 302. Richardson K, Louie-Gao Q, Arnett DK, Parnell LD, Lai CQ, Davalos A, Fox CS, Demissie S, Cupples LA, Fernandez-Hernando C, et al. The PLIN4 variant rs8887 modulates obesity related phenotypes in humans through creation of a novel mir-522 seed site. PLoS One. 2011;6:e17944. 303. Hellstrand S, Sonestedt E, Ericson U, Gullberg B, Wirfalt E, Hedblad B, Orho-Melander M. Intake levels of dietary long-chain PUFAs modify the association between genetic variation in FADS and LDL-c. J Lipid Res. 2012;53:1183–9. 304. Corella D, Sorlí J V., Estruch R, Coltell O, Ortega-Azorín C, Portolés O, Martínez-González MÁ, Bulló M, Fitó M, Arós F, et al. MicroRNA-410 regulated lipoprotein lipase variant rs13702 is associated with stroke incidence and modulated by diet in the randomized controlled PREDIMED trial. Am J Clin Nutr. 2014;100:719–31. 305. Qi Q, Durst R, Schwarzfuchs D, Leitersdorf E, Shpitzen S, Li Y, Wu H, Champagne CM, Hu FB, Stampfer MJ, et al. CETP genotype and changes in lipid levels in response to weight-loss diet intervention in the POUNDS LOST and DIRECT randomized trials. J Lipid Res. 2015;56:713–21. 306. World Health Organization. WHO (1999) Definition, Diagnosis and Classification of Diabetes Mellitus. 1999. 307. Titchenell PM, Lazar MA, Birnbaum MJ. Unraveling the regulation of hepatic metabolism by insulin. Trends Endocrinol Metab. Elsevier Ltd; 2017;28:497–505. 308. Wang CCL, Goalstone ML, Draznin B. Molecular mechanisms of insulin resistance that impace cardiovascular biology. Diabetes. 2004;53:2735–40. 309. American Diabetes Association (ADA). Standard of medical care in diabetes - 2017. Diabetes Care. 2017;40:s4– 128. 310. World Health Organization. Global report on diabetes. 2016. 311. American Diabetes Association (ADA). Diagnosis and classification of diabetes mellitus. Diabetes Care. 2010;33:s62–9. 312. Zhou B, Lu Y, Hajifathalian K, Bentham J, Di Cesare M, Danaei G, Bixby H, Cowan MJ, Ali MK, Taddei C, et al. Worldwide trends in diabetes since 1980: a pooled analysis of 751 population-based studies with 4.4 million participants. Lancet. 2016;387:1513–30. 313. International Diabetes Federation (IDF). IDF diabetes atlas seventh edition. 2015. 314. Willemsen G, Ward KJ, Bell CG, Christensen K, Bowden J, Dalgård C, Harris JR, Kaprio J, Lyle R, Magnusson PKE, et al. The concordance and heritability of type 2 diabetes in 34,166 twin pairs from international twin registers: The Discordant Twin (DISCOTWIN) Consortium. Twin Res Hum Genet. 2015;18:762–71. 315. Tallapragada DSP, Bhaskar S, Chandak GR. New insights from monogenic diabetes for “common” type 2 diabetes. Front Genet. 2015;6.

373 References

316. Skyler JS, Bakris GL, Bonifacio E, Darsow T, Eckel RH, Groop L, Groop PH, Handelsman Y, Insel RA, Mathieu C, et al. Differentiation of diabetes by pathophysiology, natural history, and prognosis. Diabetes. 2017;66:241–55. 317. Hattersley AT, Patel KA. Precision diabetes: learning from monogenic diabetes. Diabetologia. Diabetologia; 2017;60:769–77. 318. Greeley SAW, Naylor RN, Philipson LH, Bell GI. Neonatal diabetes: an expanding list of genes allows for improved diagnosis and treatment. Curr Diab Rep. 2011;11:519–32. 319. Barroso I. Genetics of type 2 diabetes. Diabet Med. 2005;22:517–35. 320. McCarthy MI. Progress in defining the molecular basis of type 2 diabetes mellitus through susceptibility-gene identification. Hum Mol Genet. 2004;13:33R–41. 321. Cauchi S, El Achhab Y, Choquet H, Dina C, Krempler F, Weitgasser R, Nejjari C, Patsch W, Chikri M, Meyre D, et al. TCF7L2 is reproducibly associated with type 2 diabetes in various ethnic groups: a global meta-analysis. J Mol Med. 2007;85:777–82. 322. Lee YH, Kang ES, Kim SH, Han SJ, Kim CH, Kim HJ, Ahn CW, Cha BS, Nam M, Nam CM, et al. Association between polymorphisms in SLC30A8, HHEX, CDKN2A/B, IGF2BP2, FTO, WFS1, CDKAL1, KCNQ1 and type 2 diabetes in the Korean population. J Hum Genet. 2008;53:991–8. 323. Ng MCY, Park KS, Oh B, Tam CHT, Cho YM, Shin HD, Lam VKL, Ma RCW, So WY, Cho YS, et al. Implication of genetic variants near TCF7L2, SLC30A8, HHEX, CDKAL1, CDKN2A/B, IGF2BP2, and FTO in type 2 diabetes and obesity in 6,719 Asians Maggie. Diabetes. 2008;57:2226–33. 324. Wu Y, Li H, Loos RJF, Yu Z, Ye X, Chen L, Pan A, Hu FB, Lin X. Common variants in CDKAL1, CDKN2A/B, IGF2BP2, SLC30A8, and HHEX/IDE genes are associated with type 2 diabetes and impaired fasting glucose in a Chinese Han population. Diabetes. 2008;57:2834–42. 325. Hu C, Zhang R, Wang C, Wang J, Ma X, Lu J, Qin W, Hou X, Wang C, Bao Y, et al. PPARG, KCNJ11, CDKAL1, CDKN2A-CDKN2B, IDE-KIF11- HHEX, IGF2BP2 and SLC30A8 are associated with type 2 diabetes in a Chinese population. PLoS One. 2009;4:e7643. 326. Rong R, Hanson RL, Ortiz D, Wiedrich C, Kobes S, Knowler WC, Bogardus C, Baier LJ. Association analysis of variation in/near FTO, CDKAL1, SLC30A8, HHEX, EXT2, IGF2BP2, LOC387761, and CDKN2B with type 2 diabetes and related quantitative traits in Pima Indians. Pima. 2009;58:478–88. 327. Chauhan G, Spurgeon CJ, Tabassum R, Bhaskar S, Kulkarni SR, Mahajan A, Chavali S, Kumar MVK, Prakash S, Dwivedi OP, et al. Impact of common variants of PPARG, KCNJ11, TCF7L2, SLC30A8, HHEX, CDKN2A, IGF2BP2, and CDKAL1 on the risk of type 2 diabetes in 5,164 Indians. Diabetes. 2010;59:2068–74. 328. Han X, Luo Y, Ren Q, Zhang X, Wang F, Sun X, Zhou X, Ji L. WFS1 in type 2 diabetes in a Chinese population. BMC Med Genet. 2010;11:10–5. 329. Dimas AS, Lagou V, Barker A, Knowles JW, Mägi R, Hivert MF, Benazzo A, Rybin D, Jackson AU, Stringham HM, et al. Impact of type 2 diabetes susceptibility variants on quantitative glycemic traits reveals mechanistic heterogeneity. Diabetes. 2014;63:2158–71. 330. Flannick J, Florez JC. Type 2 diabetes: genetic data sharing to advance complex disease research. Nat Rev Genet. 2016;17:535–49. 331. Li SX, Imamura F, Ye Z, Schulze MB, Zheng J. Interaction between genes and macronutrient intake on the risk of developing type 2 diabetes: systematic review and findings from European Prospective Investigation into Cancer (EPIC)-InterAct. Am J Clin Nutr. 2017;1–32. 332. Kanoni S, Nettleton JA, Hivert MF, Ye Z, Van Rooij FJA, Shungin D, Sonestedt E, Ngwa JS, Wojczynski MK, Lemaitre RN, et al. Total zinc intake may modify the glucose-raising effect of a zinc transporter (SLC30A8) variant: a

374 References

14-cohort meta-analysis. Diabetes. 2011;60:2407–16. 333. Fisher E, Boeing H, Fritsche A, Doering F, Joost H-G, Schulze MB. Whole-grain consumption and transcription factor-7-like 2 (TCF7L2) rs7903146: gene–diet interaction in modulating type 2 diabetes risk. Br J Nutr. 2009;101:478–81. 334. The InterAct Consortium. Investigation of gene-diet interactions in the incretin system and risk of type 2 diabetes: the EPIC-InterAct study. Diabetologia. 2016;59:2613–21. 335. Corella D, Carrasco P, Sorli J V., Estruch R, Rico-Sanz J, Martínez-González MÁ, Salas-Salvadó J, Covas MI, Coltell O, Aros F, et al. Mediterranean diet reduces the adverse effect of the TCF7L2-rs7903146 polymorphism on cardiovascular risk factors and stroke incidence. Diabetes Care. 2013;36:3803–11. 336. Cornelis MC, Qi L, Kraft P, Hu FB. TCF7L2, dietary carbohydrate, and risk of type 2 diabetes in US women. Am J ClinNutr. 2009;89:1256–62. 337. Fisher E, Schreiber S, Joost H, Boeing H, Do F. A two-step association study identifies CAV2 rs2270188 single nucleotide polymorphism interaction with fat intake in type 2 diabetes risk. J Nutr. 2011;141:177–81. 338. Hindy G, Sonestedt E, Ericson U, Jing XJ, Zhou Y, Hansson O, Renström E, Wirfält E, Orho-Melander M. Role of TCF7L2 risk variant and dietary fibre intake on incident type 2 diabetes. Diabetologia. 2012;55:2646–54. 339. Sonestedt E, Roos C, Gullberg B, Ericson U, Wirfalt E, Orho-Melander M. Fat and carbohydrate intake modify the association between genetic variation in the FTO genotype and obesity. Am J Clin Nutr. 2009;90:1418–25. 340. Wirström T, Hilding A, Gu HF, Östenson CG, Björklund A. Consumption of whole grain reduces risk of deteriorating glucose tolerance, including progression to prediabetes. Am J Clin Nutr. 2013;97:179–87. 341. Zheng JS, Huang T, Li K, Chen Y, Xie H, Xu D, Sun J, Li D. Modulation of the association between the PEPD variant and the risk of type 2 diabetes by n-3 fatty acids in Chinese Hans. J Nutrigenet Nutrigenomics. 2015;8:36–43. 342. Poulter NR, Prabhakaran D, Caulfield M. Hypertension. The Lancet. 2015. p. 801–12. 343. National Institutes of Health. National Heart Lung and Blood Institute. The seventh report of the Joint National Committee on prevention, detection, evaluation, and treatment of high blood pressure. 2004. 344. Mancia G, Fagard R, Narkiewicz K, Redon J, Zanchetti A, B??hm M, Christiaens T, Cifkova R, De Backer G, Dominiczak A, et al. 2013 ESH/ESC guidelines for the management of arterial hypertension: The Task Force for the management of arterial hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). Eur Heart J. 2013;34:2159–219. 345. World Health Organization (WHO). A global brief on hypertension. Geneva: World Health Organization; 2013. 346. Simonetti GD, Mohaupt MG, Bianchetti MG. Monogenic forms of hypertension. Eur J Pediatr. 2012;171:1433– 9. 347. Pappachan JM, Buch HN. Endocrine hypertension: a practical approach. Advs Exp Med Biol - Adv Intern Med. 2015;6:57–66. 348. Grossman A, Messerli FH, Grossman E. Drug induced hypertension - An unappreciated cause of secondary hypertension. Eur J Pharmacol. 2015;763:15–22. 349. World Health Organization (WHO). Global status report on noncommunicable diseases 2010. 2010. 350. Di Cesare M, Bentham J, Stevens GA, Zhou B, Danaei G, Lu Y, Bixby H, Cowan MJ, Riley LM, Hajifathalian K, et al. Trends in adult body-mass index in 200 countries from 1975 to 2014: A pooled analysis of 1698 population-based measurement studies with 19.2 million participants. Lancet. 2016;387:1377–96. 351. Patricia M Kearney, Megan Whelton, Kristi Reynolds, Paul Muntner, Paul K Whelton, He J. Global burden of hypertension-analysis of worldwide data. Lancet. 2005;365:217–23. 352. Frohlich ED, Dustan HP, Bumus FM. Ivine H. Page:1901-1991. Hypertension. 1991;18:443–5.

375 References

353. Kim H-K, Lee H, Kwon J-T, Kim H-J. A polymorphism in AGT and AGTR1 gene is associated with lead-related high blood pressure. J Renin-Angiotensin-Aldosterone Syst. 2015;16:712–9. 354. Vangjeli C, Clarke N, Quinn U, Dicker P, Tighe O, Ho C, O’Brien E, Stanton A V. Confirmation that the renin gene distal enhancer polymorphism REN-5312C/T is associated with increased blood pressure. Circ Cardiovasc Genet. 2010;3:53–9. 355. Zhu X, Bouzekri N, Southam L, Cooper RS, Adeyemo A, McKenzie CA, Luke A, Chen G, Elston RC, Ward R. Linkage and association analysis of angiotensin I-converting enzyme (ACE)-gene polymorphisms with ACE concentration and blood pressure. Am J Hum Genet. 2001;68:1139–48. 356. Takeuchi F, Yamamoto K, Katsuya T, Sugiyama T, Nabika T, Ohnaka K, Yamaguchi S, Takayanagi R, Ogihara T, Kato N. Reevaluation of the association of seven candidate genes with blood pressure and hypertension: a replication study and meta-analysis with a larger sample size. Hypertens Res. 2012;35:825–31. 357. Ganesh SK, Tragante V, Guo W, Guo Y, Lanktree MB, Smith EN, Johnson T, Castillo BA, Barnard J, Baumert J, et al. Loci influencing blood pressure identified using a cardiovascular gene-centric array. Hum Mol Genet. 2013;22:1663–78. 358. Burton PR, Clayton DG, Cardon LR, Craddock N, Deloukas P, Duncanson A, Kwiatkowski DP, McCarthy MI, Ouwehand WH, Samani NJ, et al. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature. 2007;447:661–78. 359. Levy D, Larson MG, Benjamin EJ, Newton-Cheh C, Wang TJ, Hwang S-J, Vasan RS, Mitchell GF. Framingham Heart Study 100K Project: genome-wide associations for blood pressure and arterial stiffness. BMC Med Genet. 2007;8:S3–14. 360. Levy D, Ehret GB, Rice K, Verwoert GC, Launer LJ, Dehghan A, Glazer NL, Morrison AC, Johnson AD, Aspelund T, et al. Genome-wide association study of blood pressure and hypertension. Nat Genet. 2009;41:677–87. 361. Newton-Cheh C, Johnson T, Gateva V, Tobin MD, Bochud M, Coin L, Najjar SS, Zhao JH, Heath SC, Eyheramendy S, et al. Genome-wide association study identifies eight loci associated with blood pressure. Nat Genet. 2009;41:666–76. 362. Ehret GB, Munroe PB, Rice KM, Bochud M, Johnson AD, Chasman DI, Smith A V., Tobin MD, Verwoert GC, Hwang S-J, et al. Genetic variants in novel pathways influence blood pressure and cardiovascular disease risk. Nature. Nature Publishing Group; 2011;478:103–9. 363. Wain L V, Verwoert GC, O’Reilly PF, Shi G, Johnson T, Johnson AD, Bochud M, Rice KM, Henneman P, Smith A V, et al. Genome-wide association study identifies six new loci influencing pulse pressure and mean arterial pressure. Nat Genet. 2011;43:1005–11. 364. Simino J, Shi G, Bis JC, Chasman DI, Ehret GB, Gu X, Guo X, Hwang SJ, Sijbrands E, Smith A V., et al. Gene-age interactions in blood pressure regulation: a large-scale investigation with the CHARGE, global BPgen, and ICBP consortia. Am J Hum Genet. 2014;95:24–38. 365. Liu C, Kraja AT, Smith JA, Brody JA, Franceschini N, Bis JC, Rice K, Morrison AC, Lu Y, Weiss S, et al. Meta- analysis identifies common and rare variants influencing blood pressure and overlapping with metabolic trait loci. Nat Genet. 2016;48:1162–70. 366. Surendran P, Drenos F, Young R, Warren H, Cook JP, Manning AK, Grarup N, Sim X, Barnes DR, Witkowska K, et al. Trans-ancestry meta-analyses identify rare and common variants associated with blood pressure and hypertension. Nat Genet. 2016;48:1151–61. 367. Kato N, Takeuchi F, Tabara Y, Kelly TN, Go MJ, Sim X, Tay WT, Chen C-H, Zhang Y, Yamamoto K, et al. Meta- analysis of genome-wide association studies identifies common variants associated with blood pressure variation in

376 References east Asians. Nat Genet. 2011;43:531–8. 368. Kelly TN, Takeuchi F, Tabara Y, Edwards TL, Kim YJ, Chen P, Li H, Wu Y, Yang CF, Zhang Y, et al. Genome-wide association study meta-analysis reveals transethnic replication of mean arterial and pulse pressure loci. Hypertension. 2013;62:853–9. 369. Hoffmann TJ, Ehret GB, Nandakumar P, Ranatunga D, Schaefer C, Kwok P-Y, Iribarren C, Chakravarti A, Risch N. Genome-wide association analyses using electronic health records identify new loci influencing blood pressure variation. Nat Genet. 2016;49:54–64. 370. Kato N, Loh M, Takeuchi F, Verweij N, Wang X, Zhang W, Kelly TN, Saleheen D, Lehne B, Leach IM, et al. Trans- ancestry genome-wide association study identifies 12 genetic loci influencing blood pressure and implicates a role for DNA methylation. Nat Genet. 2015;47:1282–93. 371. Ehret GB, Ferreira T, Chasman DI, Jackson AU, Schmidt EM, Johnson T, Thorleifsson G, Luan J, Donnelly LA, Kanoni S, et al. The genetics of blood pressure regulation and its target organs from association studies in 342,415 individuals. Nat Genet. 2016;48:1171–84. 372. Warren HR, Evangelou E, Cabrera CP, Gao H, Ren M, Mifsud B, Ntalla I, Surendran P, Liu C, Cook JP, et al. Genome-wide association analysis identifies novel blood pressure loci and offers biological insights into cardiovascular risk. Nat Genet. 2017;49. 373. Zhao Q, Kelly TN, Li C, He J. Progress and future aspects in genetics of human hypertension. Curr Hypertens Rep. 2013;15:676–86. 374. Zhao D, Qi Y, Zheng Z, Wang Y, Zhang X-Y, Li H-J, Liu H-H, Zhang X-T, Du J, Liu J. Dietary factors associated with hypertension. Nat Rev Cardiol. Nature Publishing Group; 2011;8:456–65. 375. Boutcher YN, Boutcher SH. Exercise intensity and hypertension: what’s new? J Hum Hypertens. 2017;31:157– 64. 376. Doonan RJ, Hausvater A, Scallan C, Mikhailidis DP, Pilote L, Daskalopoulou SS. The effect of smoking on arterial stiffness. Hypertens Res. Nature Publishing Group; 2010;33:398–410. 377. DeMarco VG, Aroor AR, Sowers JR. The pathophysiology of hypertension in patients with obesity. Nat Rev. 2014;10:364–76. 378. Simino J, Sung YJ, Kume R, Schwander K, Rao DC. Gene-alcohol interactions identify several novel blood pressure loci including a promising locus near SLC16A9. Front Genet. 2013;4:1–12. 379. Basson J, Sung YJ, Schwander K, Kume R, Simino J, De Las Fuentes L, Rao D. Gene-education interactions identify novel blood pressure loci in the Framingham Heart Study. Am J Hypertens. 2014;27:431–44. 380. Sung YJ, De Las Fuentes L, Schwander KL, Simino J, Rao DC. Gene-smoking interactions identify several novel blood pressure loci in the Framingham Heart Study. Am J Hypertens. 2015;28:343–54. 381. Basson J, Sung YJ, de las Fuentes L, Schwander K, Cupples LA, Rao DC. Influence of smoking status and intensity on discovery of blood pressure loci through gene-smoking interactions. Genet Epidemiol. 2015;39:277–94. 382. Taylor JY, Sun Y V, Hunt SC, Kardia SLR. Gene-environment interaction for hypertension among African American women across generations. Biol Res Nurs. 2010;12:149–55. 383. Tagetti A, Ericson U, Montagnana M, Danese E, Almgren P, Nilsson P, Engström G, Hedblad B, Minuz P, Orho- Melander M, et al. Intakes of omega-3 polyunsaturated fatty acids and blood pressure change over time: possible interaction with genes involved in 20-HETE and EETs metabolism. Prostaglandins Other Lipid Mediat. 2015;120:126– 33. 384. Li C, He J, Chen J, Zhao J, Gu D, Hixson JE, Rao DC, Jaquish CE, Gu CC, Chen J, et al. Genome-wide gene-sodium interaction analyses on blood pressure: The Genetic Epidemiology Network of Salt-Sensitivity Study. Hypertension.

377 References

2016;68:348–55. 385. Chang Y-C, Chiu Y-F, Lee I-T, Ho L-T, Hung Y-J, Hsiung CA, Quertermous T, Donlon T, Lee W-J, Lee P-C, et al. Common ALDH2 genetic variants predict development of hypertension in the SAPPHIRe prospective cohort: gene- environmental interaction with alcohol consumption. BMC Cardiovasc Disord. 2012;12:58–65. 386. Taylor JY, Schwander K, Kardia SLR, Arnett D, Liang J, Hunt SC, Rao DC, Sun Y V. A Genome-wide study of blood pressure in African Americans accounting for gene-smoking interaction. Sci Rep. 2016;6:18812. 387. Carvajal A, Centeno C, Watson R, Martínez M, Rubiales ÁS. How is an instrument for measuring health to be validated ? An Sist Sanit Navar. 2011;34:63–72. 388. Cade J, Thompson R, Burley V, Warm D. Development, validation and utilisation of food-frequency questionnaires – a review. Public Health Nutr. 2002;5:567–87. 389. Association WM. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. United States; 2013;310:2191–4. 390. Marques-Lopes I, Menal-Puey S, Martínez JA, Russolillo G. Development of a Spanish Food Exchange List: Application of Statistical Criteria to a Rationale Procedure. J Acad Nutr Diet. 2017; 391. Mataix Verdú J. Tabla de compicion de alimentos. 5th ed. Granada: Universidad de Granada; 2009. 392. Moreiras O, Carbajal Á, Cabrera L, Cuadrado C. Tablas de composición de alimentos. 16th ed. Madrid: Piramide; 2013. 393. Farrán A, Zamora R, Cervera P. Tablas de composición de alimentos del CESNID. 2nd ed. Madrid: Mc Graw Hill; 2009. 394. Department of Nutrition (UCM) & ALCE ingeniería. DIAL software for assessing diets and food calculations (for Windows, version 3.0.0.5). 2013. 395. Panel on Macronutrients Subcommittee on Upper Reference Levels of Nutrients, Subcommittee on Interpretation and Uses of Dietary Reference Intakes, and the Standing Committee on the Scientific Evaluation of Dietary Reference Intakes P on the D of DF. Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids (Macronutrients). Washington, D.C.: The National Academies Press; 2005. 396. Grim CE, Grim CM. The Omron Elite 7300W home blood pressure monitor passes the European Society of Hypertension International Validation Protocol for women and men. Blood Press Monit. 2009;14:87–90. 397. Goni L, Aray-Miranda M, Martínez JA, Cuervo M. Validation of a food groups frequency questionnaire based in an exchange system. Nutr Hosp. 2016;33:1391–9. 398. Dunbar SA. Applications of Luminex xMAP technology for rapid, high-throughput multiplexed nucleic acid detection. Clin Chim Acta. Luminex Corporation, Austin, TX 78727, USA. [email protected]; 2006;363:71– 82. 399. Mifflin MD, St Jeor ST, Hill L a, Scott BJ, Daugherty S a, Koh YO. A new predictive equation in healthy individuals for resting energy. Am J Clin Nutr. 1990;51:241–7. 400. Seagle HM, Strain GW, Makris A, Reeves RS. Position of the American Dietetic Association: weight management. J Am Diet Assoc. 2009;109:330–46. 401. Martin-Moreno JM, Boyle P, Gorgojo L, Maisonneuve P, Fernandez-Rodriguez JC, Salvini S, Willett WC. Development and validation of a food frequency questionnaire in Spain. Int J Epidemiol. 1993;22:512–9. 402. de la Fuente-Arrillaga C, Vázquez Ruiz Z, Bes-Rastrollo M, Sampson L, Martinez-González MA. Reproducibility of an FFQ validated in Spain. Public Health Nutr. 2010;13:1364–72. 403. Fernández-Ballart JD, Piñol JL, Zazpe I, Corella D, Carrasco P, Toledo E, Perez-Bauer M, Martínez-González MA,

378 References

Salas-Salvadó J, Martín-Moreno JM. Relative validity of a semi-quantitative food-frequency questionnaire in an elderly Mediterranean population of Spain. Br J Nutr. 2010;103:1808–16. 404. Martínez-González MA, López-Fontana C, Varo JJ, Sánchez-Villegas A, Martinez JA. Validation of the Spanish version of the physical activity questionnaire used in the Nurses’ Health Study and the Health Professionals’ Follow- up Study. Public Health Nutr. 2005;8:920–7. 405. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18:499–502. 406. De Souza RJ, Bray GA, Carey VJ, Hall KD, LeBoff MS, Loria CM, Laranjo NM, Sacks FM, Smith SR. Effects of 4 weight-loss diets differing in fat, protein, and carbohydrate on fat mass, lean mass, visceral adipose tissue, and hepatic fat: Results from the POUNDS LOST trial. Am J Clin Nutr. 2012;95:614–25. 407. Baecke JAH, Burema J, Frijters JER. A short questionnaire for the measurement of habitual physical activity in epidemiological studies. Am J Clin Nutr. 1982;36:936–42. 408. Levy JC, Matthews DR, Hermans MP. Correct homeostasis model assessment (HOMA) evaluation uses the computer program. Diabetes Care. 1998;21:2191–2. 409. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412–9. 410. Borai A, Livingstone C, Ferns GAA. The biochemical assessment of insulin resistance. Ann Clin Biochem. 2007;44:324–42. 411. Rohlfs R V., Weir BS. Distributions of hardy-weinberg equilibrium test statistics. Genetics. 2008;180:1609–16. 412. Prieto L, Lamarca R, Casado A. Assessment of the reliability of clinical findings: the intraclass correlation coefficient. Med Clin. 1998;110:142–5. 413. Bland JM, Altman D. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;327:307–10. 414. Lombard MJ, Steyn NP, Charlton KE, Senekal M. Application and interpretation of multiple statistical tests to evaluate validity of dietary intake assessment methods. Nutr J. Nutrition Journal; 2015;14:40–51. 415. He M, Cornelis MC, Franks PW, Zhang C, Hu FB, Qi L. Obesity genotype score and cardiovascular risk in women with type 2 diabetes mellitus. Arterioscler Thromb Vasc Biol. 2010;30:327–32. 416. Peterson RE, Maes HH, Holmans P, Sanders AR, Levinson DF, Shi J, Kendler KS, Gejman P V, Webb BT. Genetic risk sum score comprised of common polygenic variation is associated with body mass index. Hum Genet. 2011;129:221–30. 417. Cerda J, Cifuentes L. Uso de curvas ROC en investigación clínica: Aspectos teórico-prácticos. Rev Chil infectología. 2012;29:138–41. 418. Zhang H, Zamar RH. Least angle regression for model selection. WIREs Comput Stat. 2014;6:116–23. 419. Qi L. Gene-environment interactions in human diseases. New York: Nova, cop; 2016. 420. Singh M, Singh AK, Pandey P, Chandra S, Singh KA, Gambhir IS. Molecular genetics of essential hypertension. Clin Exp Hypertens. 2016;38:268–77. 421. Gerken T, Girard CA, Tung Y-CL, Webby CJ, Saudek V, Hewitson KS, Yeo GSH, McDonough MA, Cunliffe S, McNeill LA, et al. The obesity-associated FTO gene encodes a 2-oxoglutarate-dependent nucleic acid demethylase. Science. 2007;318:1469–72. 422. Zhao X, Yang Y, Sun BF, Zhao YL, Yang YG. FTO and obesity: Mechanisms of association. Curr Diab Rep. 2014;14:486–95.

379 References

423. Speakman JR. The “Fat Mass and Obesity Related” (FTO) gene: Mechanisms of impact on obesity and energy balance. Curr Obes Rep. 2015;4:73–91. 424. Rask-Andersen M, Almén M, Olausen HR, Olszewski PK, Eriksson J, Chavan RA, Levine AS, Fredriksson R, Schiöth HB. Functional coupling analysis suggests link between the obesity gene FTO and the BDNF-NTRK2 signaling pathway. BMC Neurosci. 2011;12:117. 425. Kilpeläinen TO, Carli JFM, Skowronski AA, Sun Q, Kriebel J, Feitosa MF, Hedman ÅK, Drong AW, Hayes JE, Zhao J, et al. Genome-wide meta-analysis uncovers novel loci influencing circulating leptin levels. Nat Commun. 2016;7:10494. 426. Wang P, Yang F-J. Involvement of leptin receptor long isoform (LepRb)-STAT3 signaling pathway in brain Fat Mass– and Obesity-Associated (FTO) downregulation during energy restriction. Mol Med. 2011;17:1. 427. Gulati P, Yeo GSH. The biology of FTO: From nucleic acid demethylase to amino acid sensor. Diabetologia. 2013;56:2113–21. 428. Stratigopoulos G, Martin Carli JF, O’Day DR, Wang L, Leduc CA, Lanzano P, Chung WK, Rosenbaum M, Egli D, Doherty DA, et al. Hypomorphism for RPGRIP1L, a ciliary gene vicinal to the fto locus, causes increased adiposity in mice. Cell Metab. 2014;19:767–79. 429. Smemo S, Tena JJ, Kim KH, Gamazon ER, Sakabe NJ, Gomez-Marin C, Aneas I, Credidio FL, Sobreira DR, Wasserman NF, et al. Obesity-associated variants within FTO form long-range functional connections with IRX3. Nature. 2014;507:371–5. 430. Fajas L, Auboeuf D, Raspé E, Schoonjans K, Lefebvre AM, Saladin R, Najib J, Laville M, Fruchart JC, Deeb S, et al. The organization, promoter analysis, and expression of the human PPARgamma gene. J Biol Chem. 1997;272:18779–89. 431. Robinson E, Grieve DJ. Significance of peroxisome proliferator-activated receptors in the cardiovascular system in health and disease. Pharmacol Ther. 2009;122:246–63. 432. Gouda HN, Sagoo GS, Harding AH, Yates J, Sandhu MS, Higgins JPT. The association between the peroxisome proliferator-activated receptor-γ2 (PPARG2) Pro12Ala gene variant and type 2 diabetes mellitus: A HuGE review and meta-analysis. Am J Epidemiol. 2010;171:645–55. 433. Deeb SS, Fajas L, Nemoto M, Pihlajamäki J, Mykkänen L, Kuusisto J, Laakso M, Fujimoto W, Auwerx J. A Pro12Ala substitution in PPARgamma2 associated with decreased receptor activity, lower body mass index and improved insulin sensitivity. Nat Genet. 1998;20:284–7. 434. Baldani DP, Skrgatic L, Cerne JZ, Ferk P, Simunic V, Gersak K. Association of PPARG Pro12Ala polymorphism with insulin sensitivity and body mass index in patients with polycystic ovary syndrome. Biomed Reports. 2014;2:199–206. 435. Evans D, Mann WA, de Heer J, Michel U, Wendt D, Kortner B, Wolf A, Beisiegel U. Variation in the gene for human peroxisome proliferator activated receptor gamma (PPARgamma) does not play a major role in the development of morbid obesity 2. Int J Obes Relat Metab Disord. 2000;24:647–51. 436. Hsiao TJ, Lin E. The Pro12Ala polymorphism in the peroxisome proliferator-activated receptor gamma (PPARG) gene in relation to obesity and metabolic phenotypes in a Taiwanese population. Endocrine. 2015;48:786–93. 437. Lamri A, Abi Khalil C, Jaziri R, Velho G, Lantieri O, Vol S, Froguel P, Balkau B, Marre M, Fumeron F. Dietary fat intake and polymorphisms at the PPARG locus modulate BMI and type 2 diabetes risk in the D.E.S.I.R. prospective study. Int J Obes. 2012;36:218–24. 438. Dedoussis G V, Manios Y, Kourlaba G, Kanoni S, Lagou V, Butler J, Papoutsakis C, Scott RA, Yannakoulia M, Pitsiladis YP, et al. An age-dependent diet-modified effect of the PPARgamma Pro12Ala polymorphism in children.

380 References

Metabolism. 2011;60:467–73. 439. Verier MD C, Meirhaeghe PHD A, Bokor MD, PHD S, Breidenassel PHD C, Manios PHD Y, Molnár MD, PHD D, Artero PHD EG, Nova PHD E, De Henauw MD, PHD S, Moreno MD, PHD LA, et al. Breast-feeding modulates the influence of the Peroxisome Proliferator-Activated Receptor-[gamma] (PPARG) Pro12Ala polymorphism on adiposity in adolescents. Diabetes Care. 2010;33:190–6. 440. Costa-Urrutia P, Abud C, Franco-Trecu V, Colistro V, Rodríguez-Arellano ME, Vázquez-Pérez J, Granados J, Seelaender M. Genetic obesity risk and attenuation effect of physical fitness in Mexican-Mestizo population: a case- control study. Ann Hum Genet. 2017;81:106–16. 441. Silbernagel G, Stefan N, Hoffmann MM, Machicao-Arano F, Machann J, Schick F, Winkelmann BR, Boehm BO, Häring HU, Fritsche A, et al. The L162V polymorphism of the peroxisome proliferator activated receptor alpha gene (PPARA) is not associated with type 2 diabetes, BMI or body fat composition. Exp Clin Endocrinol Diabetes. 2009;117:113–8. 442. Hatters DM, Peters-Libeu CA, Weisgraber KH. Apolipoprotein E structure: insights into function. Trends Biochem Sci. 2006;31:445–54. 443. Volcik KA, Barkley RA, Hutchinson RG, Mosley TH, Heiss G, Sharrett AR, Ballantyne CM, Boerwinkle E. Apolipoprotein E polymorphisms predict low density lipoprotein cholesterol levels and carotid artery wall thickness but not incident coronary heart disease in 12,491 ARIC study participants. Am J Epidemiol. 2006;164:342–8. 444. Zeljko HM, Škarić-Jurić T, Narančić NS, Tomas Ž, Barešić A, Salihović MP, Starčević B, Janićijević B. E2 allele of the apolipoprotein E gene polymorphism is predictive for obesity status in Roma minority population of Croatia. Lipids Health Dis. 2011;10:9–14. 445. Mazzotti DR, Singulane CC, Ota VK, Rodrigues TP, Furuya TK, de Souza FJ, Cordeiro BG, de Oliveira Amaral CM, Chen ES, Jacomini A, et al. Association of APOE, GCPII and MMP9 polymorphisms with common diseases and lipid levels in an older adult/elderly cohort. Gene. 2014;535:370–5. 446. Petkeviciene J, Smalinskiene A, Luksiene DI, Jureniene K, Ramazauskiene V, Klumbiene J, Lesauskaite V. Associations between apolipoprotein E genotype, diet, body mass index, and serum lipids in Lithuanian adult population. PLoS One. 2012;7:2–7. 447. Kypreos KE, Karagiannides I, Fotiadou EH, Karavia EA, Brinkmeier MS, Giakoumi SM, Tsompanidi EM. Mechanisms of obesity and related pathologies: role of apolipoprotein E in the development of obesity. FEBS J. 2009;276:5720–8. 448. Li Y, Liu L. Apolipoprotein E synthesized by adipocyte and apolipoprotein E carried on lipoproteins modulate adipocyte triglyceride content. Lipids Health Dis. 2014;13:136. 449. Belsky DW, Moffitt TE, Sugden K, Williams B, Houts R, McCarthy J, Caspi A. Development and evaluation of a genetic risk score for obesity. Biodemography Soc Biol. 2013;59:85–100. 450. Zhang Y, Liu J, Yao J, Ji G, Qian L, Wang J, Zhang G, Tian J, Nie Y, Zhang YE, et al. Obesity: pathophysiology and intervention. Nutrients. 2014;6:5153–83. 451. Gómez-Ambrosi J, Silva C, Galofré JC, Escalada J, Santos S, Millán D, Vila N, Ibañez P, Gil MJ, Valentí V, et al. Body mass index classification misses subjects with increased cardiometabolic risk factors related to elevated adiposity. Int J Obes. 2012;36:286–94. 452. Cascio G, Schiera G, Di Liegro I. Dietary fatty acids in metabolic syndrome, diabetes and cardiovascular diseases. Curr Diabetes Rev. 2012;8:2–17. 453. Wu L, Shen C, Seed Ahmed M, Östenson C-G, Gu HF. Adenylate cyclase 3: a new target for anti-obesity drug development. Obes Rev. 2016;17:907–14.

381 References

454. Rogne M, Taskén K. Compartmentalization of cAMP signaling in adipogenesis, lipogenesis, and lipolysis. Horm Metab Res. 2014;46:833–40. 455. Nordman S, Abulaiti A, Hilding A, Långberg E-C, Humphreys K, Ostenson C-G, Efendic S, Gu HF. Genetic variation of the adenylyl cyclase 3 (AC3) locus and its influence on type 2 diabetes and obesity susceptibility in Swedish men. Int J Obes. 2008;32:407–12. 456. Wang H, Wu M, Zhu W, Shen J, Shi X, Yang J, Zhao Q, Ni C, Xu Y, Shen H, et al. Evaluation of the association between the AC3 genetic polymorphisms and obesity in a Chinese Han population. PLoS One. 2010;5:e13851. 457. Wang Z, Li V, Chan GCK, Phan T, Nudelman AS, Xia Z, Storm DR. Adult type 3 adenylyl cyclase-deficient mice are obese. PLoS One. 2009;4:e6979. 458. Cao H, Chen X, Yang Y, Storm DR. Disruption of type 3 adenylyl cyclase expression in the hypothalamus leads to obesity. Integr Obes Diabetes. 2016;28:1304–14. 459. Seed Ahmed M, Kovoor A, Nordman S, Abu Seman N, Gu T, Efendic S, Brismar K, Östenson C-G, Gu HF. Increased expression of adenylyl cyclase 3 in pancreatic islets and central nervous system of diabetic Goto-Kakizaki rats: a possible regulatory role in glucose homeostasis. Islets. 2012;4:343–8. 460. Chen X, Luo J, Leng Y, Yang Y, Zweifel LS, Palmiter RD, Storm DR. Ablation of type III adenylyl cyclase in mice causes reduced neuronal activity, altered sleep pattern, and depression-like phenotypes. Biol Psychiatry. 2016;80:836–48. 461. Qiu L, LeBel RP, Storm DR, Chen X. Type 3 adenylyl cyclase: A key enzyme mediating the cAMP signaling in neuronal cilia. Int J Physiol Pathophysiol Pharmacol. 2016;8:95–108. 462. You P, Hu H, Chen Y, Zhao Y, Yang Y, Wang T, Xing R, Shao Y, Zhang W, Li D, et al. Effects of melanocortin 3 and 4 receptor deficiency on energy homeostasis in rats. Sci Rep. 2016;6:34938. 463. Tong T, Shen Y, Lee H-W, Yu R, Park T. Adenylyl cyclase 3 haploinsufficiency confers susceptibility to diet- induced obesity and insulin resistance in mice. Sci Rep. 2016;6:34179. 464. Pitman JL, Wheeler MC, Lloyd DJ, Walker JR, Glynne RJ, Gekakis N. A gain-of-function mutation in adenylate cyclase 3 protects mice from diet-induced obesity. PLoS One. 2014;9:e110226. 465. Laermans J, Depoortere I. Chronobesity: Role of the circadian system in the obesity epidemic. Obes Rev. 2016;17:108–25. 466. Garaulet M, Ordovás JM, Madrid JA. The chronobiology, etiology and pathophysiology of obesity. Int J Obes. 2010;34:1667–83. 467. Qian J, Scheer FAJL. Circadian system and glucose metabolism: Implications for physiology and disease. Trends Endocrinol Metab. 2016;27:282–93. 468. Scott EM. Circadian clocks, obesity and cardiometabolic function. Diabetes, Obes Metab. 2015;17:84–9. 469. Cipolla-Neto J, Amaral FG, Afeche SC, Tan DX, Reiter RJ. Melatonin, energy metabolism, and obesity: a review. J Pineal Res. 2014;56:371–81. 470. Dubocovich ML, Delagrange P, Krause DN, Sugden D, Cardinali DP, Olcese J. Nomenclature, classification, and pharmacology of G Protein-Coupled Melatonin Receptors. Pharmocological Rev. 2010;62:343–80. 471. Gamble KL, Berry R, Frank SJ, Young ME. Circadian clock control of endocrine factors. Nat Rev Endocrinol. 2014;10:466–75. 472. Prokopenko I, Langenberg C, Florez JC, Saxena R, Soranzo N, Thorleifsson G, Loos RJF, Manning AK, Jackson AU, Aulchenko Y, et al. Variants in MTNR1B influence fasting glucose levels. Nat Genet. 2009;41:77–81. 473. Manning AK, Hivert MF, Scott RA, Grimsby JL, Bouatia-Naji N, Chen H, Rybin D, Liu CT, Bielak LF, Prokopenko I, et al. A genome-wide approach accounting for body mass index identifies genetic variants influencing fasting

382 References glycemic traits and insulin resistance. Nat Genet. 2012;44:659–69. 474. Gaulton KJ, Ferreira T, Lee Y, Raimondo A, Mägi R, Reschen ME, Mahajan A, Locke A, William Rayner N, Robertson N, et al. Genetic fine mapping and genomic annotation defines causal mechanisms at type 2 diabetes susceptibility loci. Nat Genet. 2015;47:1415–25. 475. Stancáková A, Kuulasmaa T, Paananen J, Jackson AU, Bonnycastle LL, Collins FS, Boehnke M, Kuusisto J, Laakso M. Association of 18 confirmed susceptibility loci for type 2 diabetes with indices of insulin release, proinsulin conversion, and insulin sensitivity in 5,327 nondiabetic Finnish men. Diabetes. 2009;58:2129–36. 476. Kong X, Zhang X, Xing X, Zhang B, Hong J, Yang W. The association of type 2 diabetes loci identified in genome- wide association studies with metabolic syndrome and its components in a Chinese population with type 2 diabetes. PLoS One. 2015;10:1–21. 477. Grotenfelt NE, Wasenius NS, Rönö K, Laivuori H, Stach-Lempinen B, Orho-Melander M, Schulz CA, Kautiainen H, Koivusalo SB, Eriksson JG. Interaction between rs10830963 polymorphism in MTNR1B and lifestyle intervention on occurrence of gestational diabetes. Diabetologia. Diabetologia; 2016;59:1655–8. 478. Tuomi T, Nagorny CLF, Singh P, Bennet H, Yu Q, Alenkvist I, Isomaa B, Östman B, Söderström J, Pesonen AK, et al. Increased melatonin signaling is a risk factor for type 2 diabetes. Cell Metab. 2016;23:1067–77. 479. Kumar Jha P, Challet E, Kalsbeek A. Circadian rhythms in glucose and lipid metabolism in nocturnal and diurnal mammals. Mol Cell Endocrinol. 2015;418:74–88. 480. Wolden-Hanson T, Mitton DR, McCants RL, Yellon SM, Wilkinson CW, Matsumoto AM, Rasmussen DD. Daily melatonin administration to middle-aged male rats suppresses body weight, intraabdominal adiposity, and plasma leptin and insulin independent of food intake and total body fat. Endocrinology. 2000;141:487–97. 481. Srinivasan V, Ohta Y, Espino J, Pariente JA, Rodriguez AB, Mohamed M, Zakaria R. Metabolic syndrome, its pathophysiology and the role of melatonin. 2013;7:11–25. 482. Lane JM, Chang AM, Bjonnes AC, Aeschbach D, Anderson C, Cade BE, Cain SW, Czeisler CA, Gharib SA, Gooley JJ, et al. Impact of common diabetes risk variant in MTNR1B on sleep, circadian, and melatonin physiology. Diabetes. 2016;65:1741–51. 483. Oishi K, Uchida D, Itoh N. Low-carbohydrate, high-protein diet affects rhythmic expression of gluconeogenic regulatory and circadian clock genes in mouse peripheral tissues. Chronobiol Int. 2012;29:799–809. 484. Oosterman JE, Kalsbeek A, la Fleur SE, Belsham DD. Impact of nutrients on circadian rhythmicity. Am J Physiol 2015;308:R337–50. 485. Yanagihara H, Ando H, Hayashi Y, Obi Y, Fujimura A. High-fat feeding exerts minimal effects on rhythmic mRNA expression of clock genes in mouse peripheral tissues. Chronobiol Int. 2006;23:905–14. 486. Hsieh M-C, Yang S-C, Tseng H-L, Hwang L-L, Chen C-T, Shieh K-R. Abnormal expressions of circadian-clock and circadian clock-controlled genes in the livers and kidneys of long-term, high-fat-diet-treated mice. Int J Obes (Lond). 2010;34:227–39. 487. Cano P, Jiménez-Ortega V, Larrad Á, Toso CFR, Cardinali DP, Esquifino AI. Effect of a high-fat diet on 24-h pattern of circulating levels of prolactin, luteinizing hormone, testosterone, corticosterone, thyroid-stimulating hormone and glucose, and pineal melatonin content, in rats. Endocrine. 2008;33:118–25. 488. Cano P, Cardinali DP, Rios-Lugo MJ, Fernandez-Mateos MP, Reyes Toso CF, Esquifino AI. Effect of a high-fat diet on 24-hour pattern of circulating adipocytokines in rats. Obesity (Silver Spring). 2009;17:1866–71. 489. Cardinali DP, Cano P, Jiménez-Ortega V, Esquifino AI. Melatonin and the metabolic syndrome: Physiopathologic and therapeutical implications. Neuroendocrinology. 2011;93:133–42. 490. Gooley JJ. Circadian regulation of lipid metabolism. Proc Nutr Soc. 2016;75:440–50.

383 References

491. Sun H, Huang F, Qu S. Melatonin: a potential intervention for hepatic steatosis. Lipids Health Dis. 2015;14:75– 80. 492. Koziróg M, Poliwczak AR, Duchnowicz P, Koter-Michalak M, Sikora J, Broncel M. Melatonin treatment improves blood pressure, lipid profile, and parameters of oxidative stress in patients with metabolic syndrome. J Pineal Res. 2011;50:261–6. 493. Kadhim HM, Ismail SH, Hussein KI, Bakir IH, Sahib AS, Khalaf BH, Hussain SAR. Effects of melatonin and zinc on lipid profile and renal function in type 2 diabetic patients poorly controlled with metformin. J Pineal Res. 2006;41:189–93. 494. DeMenna J, Puppala S, Chittoor G, Schneider J, Kim JY, Shaibi GQ, Mandarino LJ, Duggirala R, Coletta DK. Association of common genetic variants with diabetes and metabolic syndrome related traits in the Arizona insulin resistance registry: A focus on Mexican American families in the southwest. Hum Hered. 2014;78:47–58. 495. Sun L, Wang Y, Song Y, Cheng XR, Xia S, Rahman MRT, Shi Y, Le G. Resveratrol restores the circadian rhythmic disorder of lipid metabolism induced by high-fat diet in mice. Biochem Biophys Res Commun. 2015;458:86–91. 496. Belsky J, Jonassaint C, Pluess M, Stanton M, Brummett B, Williams R. Vulnerability genes or plasticity genes? Mol Psychiatry. 2009;14:746–54. 497. Belsky J, Hartman S. Gene-environment interaction in evolutionary perspective: differential susceptibility to environmental influences. World Psychiatry. 2014;13:87–9. 498. Hartman S, Belsky J. An evolutionary perspective on family studies: differential susceptibility to environmental influences. Fam Process. 2016;55:1–13. 499. Dalle Molle R, Fatemi H, Dagher A, Levitan RD, Silveira PP, Dubé L. Gene and environment interaction: is the differential susceptibility hypothesis relevant for obesity? Neurosci Biobehav Rev. 2016;73:326–39. 500. Rauschert S, Uhl O, Koletzko B, Hellmuth C. Metabolomic biomarkers for obesity in humans: A short review. Ann Nutr Metab. 2014;64:314–24. 501. Guasch-Ferré M, Hruby A, Toledo E, Clish CB, Martínez-González MA, Salas-Salvadó J, Hu FB. Metabolomics in prediabetes and diabetes: A systematic review and meta-analysis. Diabetes Care. 2016;39:833–46. 502. Yoon M-S. The Emerging Role of Branched-Chain Amino Acids in Insulin Resistance and Metabolism. Nutrients. 2016;8:405. 503. Lynch CJ, Adams SH. Branched-chain amino acids in metabolic signalling and insulin resistance. Nat Rev Endocrinol. 2014;10:723–36. 504. Lotta LA, Scott RA, Sharp SJ, Burgess S, Luan J, Tillin T, Schmidt AF, Imamura F, Stewart ID, Perry JRB, et al. Genetic predisposition to an impaired metabolism of the Branched-Chain Amino Acids and risk of Type 2 Diabetes: A Mendelian randomisation analysis. PLoS Med. 2016;13:1–22. 505. Mahendran Y, Jonsson A, Have CT, Allin KH, Witte DR, Jørgensen ME, Grarup N, Pedersen O, Kilpeläinen TO, Hansen T. Genetic evidence of a causal effect of insulin resistance on branched-chain amino acid levels. Diabetologia. 2017;60:873–8. 506. Gieger C, Geistlinger L, Altmaier E, De Angelis MH, Kronenberg F, Meitinger T, Mewes HW, Wichmann HE, Weinberger KM, Adamski J, et al. Genetics meets metabolomics: A genome-wide association study of metabolite profiles in human serum. PLoS Genet. 2008;4:e1000282. 507. Kettunen J, Tukiainen T, Sarin A-P, Ortega-Alonso A, Tikkanen E, Lyytikäinen L-P, Kangas AJ, Soininen P, Würtz P, Silander K, et al. Genome-wide association study identifies multiple loci influencing human serum metabolite levels. Nat Genet. 2012;44:269–76. 508. Rhee EP, Ho JE, Chen MH, Shen D, Cheng S, Larson MG, Ghorbani A, Shi X, Helenius IT, O’Donnell CJ, et al. A

384 References genome-wide association study of the human metabolome in a community-based cohort. Cell Metab. 2013;18:130– 43. 509. Shin S-Y, Fauman EB, Petersen A-K, Krumsiek J, Santos R, Huang J, Arnold M, Erte I, Forgetta V, Yang T-P, et al. An atlas of genetic influences on human blood metabolites. Nat Genet. 2014;46:543–50. 510. Draisma HHM, Pool R, Kobl M, Jansen R, Petersen A-K, Vaarhorst AAM, Yet I, Haller T, Demirkan A, Esko T, et al. Genome-wide association study identifies novel genetic variants contributing to variation in blood metabolite levels. Nat Commun. 2015;6:7208. 511. Kettunen J, Demirkan A, Würtz P, Draisma HHMM, Haller T, Rawal R, Vaarhorst A, Kangas AJ, Lyytikäinen L-P, Pirinen M, et al. Genome-wide study for circulating metabolites identifies 62 loci and reveals novel systemic effects of LPA. Nat Commun. 2016;7:11122. 512. Joshi M, Jeoung NH, Popov KM, Harris RA. Identification of a novel PP2C-type mitochondrial phosphatase. Biochem Biophys Res Commun. 2007;356:38–44. 513. Lu G, Sun H, She P, Youn J, Warburton S, Ping P, Vondriska TM, Cai H, Lynch CJ, Wang Y. Protein phosphatase 2Cm is a critical regulator of branched-chain amino acid catabolism in mice and cultured cells. J Clin Invest. 2009;119:1678–87. 514. Soeters P, Fischer J. Insulin, glucagon, aminoacid imbalance, and hepatic encephalopathy. Lancet (London, England). 1976;2:880–2. 515. Taneera J, Lang S, Sharma A, Fadista J, Zhou Y, Ahlqvist E, Jonsson A, Lyssenko V, Vikman P, Hansson O, et al. A systems genetics approach identifies genes and pathways for type 2 diabetes in human islets. Cell Metab. 2012;16:122–34. 516. Oyarzabal A, Martínez-Pardo M, Merinero B, Navarrete R, Desviat LR, Ugarte M, Rodríguez-Pombo P. A novel regulatory defect in the Branched-Chain alpha-Keto Acid Dehydrogenase Complex due to a mutation in the PPM1K gene causes a mild variant phenotype of maple syrup urine disease. Hum Mutat. 2013;34:355–62. 517. Loh P-R, Bhatia G, Gusev A, Finucane HK, Bulik-Sullivan BK, Pollack SJ, Psychiatric Genomics Consortium SWG, de Candia TR, Lee SH, Wray NR, et al. Contrasting regional architectures of schizophrenia and other complex diseases using fast variance components analysis. Nat Genet. 2015;47:16527. 518. Zimmerman HA, Olson KC, Chen G, Lynch CJ. Adipose transplant for inborn errors of branched chain amino acid metabolism in mice. Mol Genet Metab. 2013;109:345–53. 519. Newgard CB, An J, Bain JR, Muehlbauer MJ, Stevens RD, Lien LF, Haqq AM, Shah SH, Arlotto M, Slentz CA, et al. A Branched-Chain Amino Acid-related metabolic signature that differentiates obese and lean humans and contributes to insulin resistance. Cell Metab. 2009;9:311–26. 520. Newgard CB. Interplay between lipids and branched-chain amino acids in development of insulin resistance. Cell Metab. 2012;15:606–14. 521. Kadota Y, Toyoda T, Kitaura Y, Adams SH, Shimomura Y. Regulation of hepatic branched-chain α-ketoacid dehydrogenase complex in rats fed a high-fat diet. Obes Res Clin Pract. 2013;7:e439–44. 522. Shaham O, Wei R, Wang TJ, Ricciardi C, Lewis GD, Vasan RS, Carr SA, Thadhani R, Gerszten RE, Mootha VK. Metabolic profiling of the human response to a glucose challenge reveals distinct axes of insulin sensitivity. Mol Syst Biol. 2008;4:1–9. 523. Lian K, Du C, Liu Y, Zhu D, Yan W, Zhang H, Hong Z, Liu P, Zhang L, Pei H, et al. Impaired adiponectin signaling contributes to disturbed catabolism of branched-chain amino acids in diabetic mice. Diabetes. 2015;64:49–59. 524. Ho JE, Larson MG, Vasan RS, Ghorbani A, Cheng S, Rhee EP, Florez JC, Clish CB, Gerszten RE, Wang TJ. Metabolite profiles during oral glucose challenge. Diabetes. 2013;62:2689–98.

385 References

525. Butler MG. Genetics of hypertension. Current status. J Med Liban. 2010;58:175–8. 526. Gamboa A, Shibao C, Diedrich A, Choi L, Pohar B, Jordan J, Paranjape S, Farley G, Biaggioni I. Contribution of endothelial nitric oxide to blood pressure in humans. Hypertension. 2007;49:170–7. 527. Niu W, Qi Y. An updated meta-analysis of endothelial nitric oxide synthase gene: three well-characterized polymorphisms with hypertension. PLoS One. 2011;6:e24266. 528. Conen D, Cheng S, Steiner LL, Buring JE, Ridker PM, Zee RY. Association of 77 polymorphisms in 52 candidate genes with blood pressure progression and incident hypertension: the Women’s Genome Health Study. J Hypertens. 2009;27:476–83. 529. Persu A, Stoenoiu MS, Messiaen T, Davila S, Robino C, El-Khattabi O, Mourad M, Horie S, Feron O, Balligand JL, et al. Modifier effect of ENOS in autosomal dominant polycystic kidney disease. Hum Mol Genet. 2002;11:229–41. 530. Tesauro M, Thompson WC, Rogliani P, Qi L, Chaudhary PP, Moss J. Intracellular processing of endothelial nitric oxide synthase isoforms associated with differences in severity of cardiopulmonary diseases: cleavage of proteins with aspartate vs. glutamate at position 298. Proc Natl Acad Sci USA. 2000;97:2832–5. 531. Newton-Cheh C, Johnson T, Gateva V, Tobin MD, Bochud M, Coin L, Najjar SS, Zhao JH, Heath SC, Eyheramendy S, et al. Genome-wide association study identifies eight loci associated with blood pressure. Nat Genet. 2009;41:666–76. 532. Fairchild TA, Fulton D, Fontana JT, Gratton JP, McCabe TJ, Sessa WC. Acidic hydrolysis as a mechanism for the cleavage of the Glu(298)-->Asp variant of human endothelial nitric-oxide synthase. J Biol Chem. 2001;276:26674–9. 533. McDonald DM, Alp NJ, Channon KM. Functional comparison of the endothelial nitric oxide synthase Glu298Asp polymorphic variants in human endothelial cells. Pharmacogenetics. 2004;14:831–9. 534. Joshi MS, Mineo C, Shaul PW, Bauer JA. Biochemical consequences of the NOS3 Glu298Asp variation in human endothelium: altered caveolar localization and impaired response to shear. FASEB J. 2007;21:2655–63. 535. Kingah PL, Luu HN, Volcik KA, Morrison AC, Nettleton JA, Boerwinkle E. Association of NOS3 Glu298Asp SNP with hypertension and possible effect modification of dietary fat intake in the ARIC study. Hypertens Res. 2010;33:165–9. 536. Pereira T V, Rudnicki M, Cheung BM, Baum L, Yamada Y, Oliveira PS, Pereira AC, Krieger JE. Three endothelial nitric oxide (NOS3) gene polymorphisms in hypertensive and normotensive individuals: meta-analysis of 53 studies reveals evidence of publication bias. J Hypertens. 2007;25:1763–74. 537. Martins MA, Catta-Preta M, Mandarim-de-Lacerda CA, Aguila MB, Brunini TC, Mendes-Ribeiro AC. High fat diets modulate nitric oxide biosynthesis and antioxidant defence in red blood cells from C57BL/6 mice. Arch Biochem Biophys. 2010;499:56–61. 538. Abdel-Aziz TA, Mohamed RH. Association of endothelial nitric oxide synthase gene polymorphisms with classical risk factors in development of premature coronary artery disease. Mol Biol Rep. 2013;40:3065–71. 539. Silva DA, Petroski EL, Peres MA. Accuracy and measures of association of anthropometric indexes of obesity to identify the presence of hypertension in adults: a population-based study in Southern Brazil. Eur J Nutr. 2013;52:237–46. 540. Hickner RC, Kemeny G, Stallings HW, Manning SM, McIver KL. Relationship between body composition and skeletal muscle eNOS. Int J Obes (Lond). 2006;30:308–12. 541. Kraus RM, Houmard JA, Kraus WE, Tanner CJ, Pierce JR, Choi MD, Hickner RC. Obesity, insulin resistance, and skeletal muscle nitric oxide synthase. J Appl Physiol, 2012;113:758–65. 542. da Cunha N V, Pinge-Filho P, Panis C, Silva BR, Pernomian L, Grando MD, Cecchini R, Bendhack LM, Martins- Pinge MC. Decreased endothelial nitric oxide, systemic oxidative stress, and increased sympathetic modulation

386 References contribute to hypertension in obese rats. Am J Physiol Circ Physiol. 2014;306:H1472-80. 543. Smith RA, Levine TR, Lachlan KA, Fediuk TA. The high cost of complexity in experimental design and data analysis: type I and type II error rates in multiway ANOVA. Hum Commun Res. 2002;25:515–30. 544. Cohen J, Cohen P, West S, Aiken L. Applied multiple regression/correlation analysis for the behavioral sciences. 3rd ed. Cohen J, Cohen P, West S, Aiken L, editors. New Jersey: Lawrence Erlbaum Associates; 2013. 545. Keppel G, Wickens T. Design and analysis: a researcher’s handbook. 4th ed. Keppel G, Wickens T, editors. Pearson Prentice Hall; 2004. 546. Qi L. Mendelian randomization in nutritional epidemiology. Nutr Rev. 2009;67:439–50.

387