Epistatic Interactions and Epigenetic Modifications for Molecular Stratification of Chronic Diseases Ting Xie

Epistatic interactions and epigenetic modifications for molecular stratification of chronic diseases Ting Xie

To cite this version:

Ting Xie. Epistatic interactions and epigenetic modifications for molecular stratification of chronic diseases. Human health and pathology. Université de Lorraine, 2017. English. NNT : 2017LORR0339. tel-01835056

HAL Id: tel-01835056 https://tel.archives-ouvertes.fr/tel-01835056 Submitted on 11 Jul 2018

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés.

AVERTISSEMENT

Ce document est le fruit d'un long travail approuvé par le jury de soutenance et mis à disposition de l'ensemble de la communauté universitaire élargie.

Il est soumis à la propriété intellectuelle de l'auteur. Ceci implique une obligation de citation et de référencement lors de l’utilisation de ce document.

D'autre part, toute contrefaçon, plagiat, reproduction illicite encourt une poursuite pénale.

Contact : [email protected]

LIENS

Code de la Propriété Intellectuelle. articles L 122. 4 Code de la Propriété Intellectuelle. articles L 335.2- L 335.10 http://www.cfcopies.com/V2/leg/leg_droi.php http://www.culture.gouv.fr/culture/infos-pratiques/droits/protection.htm

Ecole Doctorale BioSE (Biologie- Santé- Environnement)

Thèse

Présentée et soutenue publiquement pour l’obtention du titre de

DOCTEUR DE l’UNIVERSITE DE LORRAINE

Mention: « Sciences de la Vie et de la Santé »

Par: Ting XIE

Interactions épistatiques et modifications épigénétiques pour la stratification moléculaire des maladies chroniques Epistatic interactions and epigenetic modifications for molecular stratification of chronic diseases.

Le 19 Décembre 2017 Membres du jury: Rapporteurs : Prof Nikolaos DRAKOULIS National & Kapodistrian University of Athens, Faculty of Pharmacy, Athens Prof. Dr. Belgin SUSLEYICI Marmara University, Faculty of Science and Letters, Department of Molecular Biology, Istanbul Examinateurs : Prof Jochen SCHNEIDER Luxembourg Centre for Systems Biomedicine - Université du Luxembourg, Luxembourg Prof Jean-Louis MERLIN Faculté de Pharmacie, Université de Lorraine, Nancy Dr Maria STATHOPOULOU IR2, INSERM U1122; IGE-PCV, Université de Lorraine, Nancy (co-directrice de thèse) Dr Sofia SIEST DR1, INSERM U1122; IGE-PCV, Université de Lorraine, Nancy (directrice de thèse)

Université de Lorraine - UMR INSERM U1122; IGE-PCV - Faculté de Pharmacie - 30, rue Lionnois - 54000 Nancy – France 1

Acknowledgement Many persons contributed to this thesis in different ways, and I am grateful to all of them.

My biggest thank is for my major supervisor, Dr Sophie VISVIKIS-SIEST, Director of the UMR INSERM U1122; IGE-PCV, who gave me the opportunity to work with her and her team, and gave me a very exciting project. Without her help and support, this project would not have been possible.

I would like also to thank my co-supervisor, Dr Maria STATHOPOULOU for being helpful for correction of my thesis manuscript, and patiently explained the scientific problems with her rigorous scientific spirit during my research.

I would like to express my sincere thanks to members of my thesis committee, Prof Nikolaos DRAKOULIS, Prof Belgin SÜSLEYİCİ DUMAN, Prof Jochen SCHNEIDER and Dr Jean-Louis MERLIN for having accepted to evaluate my Ph.D.

I would like to thank all of my colleagues in U1122; IGE-PCV team, present and past: Brigitte, Patricia, Christine, Sébastien, Alex-Ander, Alexandros, Jerome, Vesna, Samina, Dwaine and Lauren.

I would like to thank the ‘Region Lorraine’ for having awarded me a three-year scholarship, which enabled me to carry out this project. I would like also to thank Randox Laboratories Ltd, the "Agence Nationale de la Recherche, programme d'Investissements d'avenir"(grant number ANR-15RHU-0004) and the Operational Programme “FEDER-FSE Lorraine et Massif des Vosges 2014-2020” for their help in the realisation of the projects

And finally, I would like to extend my deepest gratitude to my family: my parents Aiguo XIE and Lianyu SUN, and so on. They always have provided unwavering love and encouragement. Thank you for believing in me.

Table of contents

Acknowledgement ...... 2

Abbreviations ...... 6

List of tables ...... 9

List of figures ...... 11

List of publications ...... 12

List of original publications ...... 13

List of review publications ...... 14

Other publications ...... 14

Présentation synthétique de la thèse ...... 16

Situation du sujet ...... 17

Résultats obtenus - discussion ...... 20

Conclusions et perspectives ...... 28

Foreword ...... 30

Introduction ...... 33

Chronic diseases ...... 34

Cardiovascular diseases ...... 35

The major CVD risk factors ...... 35

Age and gender ...... 36

Family inheritance ...... 36

Tobacco ...... 37

Hypertension ...... 39

Obesity ...... 39

Dyslipidemia and blood lipids ...... 40

Major depressive disorder (MDD) ...... 43

The risk factors for MDD ...... 44

Osteoporosis ...... 45

Alzheimer’s disease AD ...... 46 Common biomarkers ...... 47

Comorbidities ...... 48

Epistasis ...... 50

Epigenetics ...... 51

The targeted candidate biomarkers ...... 53

Vascular endothelial growth factor (VEGF) ...... 54

Apolipoprotein E (ApoE) ...... 58

Lipolysis-stimulated lipoprotein receptor (LSR) ...... 58

Personalised medicine ...... 59

Pharmacogenomics ...... 60

Materials and Methods ...... 66

Populations ...... 67

STANISLAS Family Study ...... 67

Cilento study ...... 67

LifeLines Cohort Study and Biobank (LLs) ...... 68

ApoEurope cohort ...... 68

Hypertensive population...... 69

Case-control study of depression ...... 69

Populations of postmenopausal women: ...... 70

Alzheimer’s case-control population ...... 70 Genotyping – genome-scale methylation assay ...... 70 Statistical analysis ...... 73

Publications - Results ...... 74

Publication 1 ...... 75

Publication 2 ...... 106

Publication 3 ...... 114

Publication 4 ...... 119

Publication 5 ...... 129

Publication 6 ...... 158

rs34259399 x ...... 161 APOE non-e4 allele...... 161 4

rs916147 x ...... 161 APOE non-e4 allele ...... 161 Publication 7 ...... 177

Publication 8 ...... 201

General Discussion ...... 222

General conclusion ...... 225

References ...... 227

Abstract ...... 249

Résumé ...... 249

Abbreviations AD...... Alzheimer's disease

APP…………...Amyloid precursor protein

Aβ…………….Amyloid beta

APOB...... Apolipoprotein B

APOE...... Apolipoprotein E

ARVC/D………Arrhythmogenic right ventricular cardiomyopathy / dysplasia

BMD...... Bone mineral density

BMI……….….Body mass index

BRC………….Biological Resources Center

CAD...... Coronary artery disease

CDs...... Chronic diseases

CVD...... Cardiovascular disease

CHD...... Coronary heart disease

COPD…….…..Chronic obstructive pulmonary disease

CRP...... C-reactive protein

CNS...... Central nervous system

DKK1…………Dickkopf 1

DNMT………...DNA methyltransferase

EWAS...... Epigenome-wide association study

GWAS...... Genome-Wide Association Studies

HCM…………..Familial hypertrophic cardiomyopathy

HDL...... High-density lipoprotein

HLP…………...Hyperlipidemia

IL6...... Interleukin-6

KREMEN2……Kringle Containing Transmembrane Protein 2

LAMP2………..Lysosomal-associated membrane protein 2

LAP…………...Lipid accumulation product

LDL...... Low-density lipoprotein

LDL-R...... Low density lipoprotein-receptor

LLs…………….LifeLines Cohort

LRP6…………..Low-density lipoprotein-related receptor 6

LSR...... Lipolysis-stimulated receptor

MAF...... Minor Allele Frequency

MDD...... Major depressive disorder mDNA...... DNA methylation

MZF1………….Myeloid zinc finger 1

PRKAG2……...Protein kinase AMP-activated non-catalytic subunit gamma 2

PGx……………Pharmacogenomics

SNPs…………..Single Nucleotide Polymorphisms

SD……………..Standard Deviation

SE……………...Standard error

SFS…………….STANISLAS Family Study

T2D……………Type 2 diabetes

TMEM150A…...Transmembrane Protein 150A

TNF-α...... Tumor necrosis factor-α

TSS……………Transcription start site

ST8SIA5………Alpha-2,8-sialyltransferase 8E

TG...... Triglycerides 7

TC...... Total cholesterol

VAI…………….Vascular adiposity index

VEGF...... Vascular endothelial growth factor

WC...... Waist circumference

WDR83OS…….WD Repeat Domain 83 Opposite Strand

List of tables

Table 1 Classification of overweight and obesity by percentage of body 40 fat, body mass index, waist circumference (WC) and associated diseases risk Table 2 Size of mRNA and protein of VEGF family members 55 Table 3 Characteristics of the assessed VEGF, LSR and APOE 72 polymorphisms Table 4 Significant epistatic interactions on CVD intermediate 109 phenotypes in the VEGF Consortium meta-analysis Table 5 Characteristics of post-menopausal women populations 122 Table 6 Polymorphisms’ characteristics in the post-menopausal women 123 populations Table 7A Associations of epistatic interactions with FnBMD 124 Table 7B Meta-analysis of FnBMD results 125 Table 8A Associations of interactions with age of menopause 126 Table 8B Meta-analysis of age of menopause results 127 Table 9 Characteristics of the SFS adults and children and the ApoEurope 132 replication population Table 10 Polymorphisms’ characteristics of the SFS adults and children and 133 the ApoEurope replication population Table 11 Direct effects of rs916147 LSR gene polymorphism 134 Table 12A Significant epistatic interactions of rs916147 SNP of LSR with 135 APOE polymorphism Table 12B Direct effect of APOE genotypes on the above lipids traits 136 Table 13 AD patients and controls characteristics 160 Table 14 Polymorphisms’ characteristics of AD patients and controls 160 Table 15 Significant epistatic interactions of rs34259399 and rs916147 161

SNPs of LSR with APOE Table 16 Significant epistatic interactions between APOE and rs34259399 161 Table 17 Significant epistatic interactions between APOE and rs916147 161 Table 18 Characteristics of the SFS population 179 Table 19 Association of methylation site with waist circumference level 179 Table 20 Characteristics of the SFS population 204 Table 21 Association of methylation values with TG level 205

List of figures Figure 1 10 common chronic diseases for adults >65 years old 34 Figure 2 The potential mechanism of cigarette smoking initiates 38 cardiovascular dysfunction Figure 3 Pathophysiology of atherosclerosis through cholesterol 42 synthesis Figure 4 MDD influences various diseases 44 Figure 5 Neural systems of relevance to major depressive disorder 45 Figure 6 Comorbidity among CVD, arthritis, back pain and mental 49 health conditions Figure 7 Comorbidities among selected chronic diseases by age 50 Figure 8 DNA methylation, the methylation of cytosine from 52 5-methylcytosine Figure 9 The most common histone modifications 53 Figure 10 VEGF receptor-binding properties and co-receptors 55 Figure 11 Structure of the main splicing variants of VEGFA in humans 57 Figure 12 The effect of rs6921438 x rs6993770 on FNBMD could be 128 mediated by IL6 Figure 13 cg16170243 could modify the aberrant transcription of 180 AC090241.2 locus and repress its expression, which can lead to an antisense transcript that overlaps with ST8SIA5 gene Figure 14 Summary of the identified links between the selected chronic 226 diseases and the targeted markers through investigation of epistatic interactions and mDNA

List of publications

List of original publications

1. Stathopoulou MG, Xie T, Ruggiero D, Chatelin J, Rancier M, Weryha G, Kurth MJ, Aldasoro Arguinano AA, Gorenjak V, Petrelis AM, Dagher G, Dedoussis G, Deloukas P, Lamont J, Marc J, Simmaco M, van Schaik RHN, Innocenti F, Merlin JL, Schneider J, Alizadeh BZ, Ciullo M, Seshadri S, Visvikis-Siest S, VEGF Consortium. A transnational collaborative network dedicated to the study and applications of the Vascular Endothelial Growth Factor–A in medical practice: The VEGF Consortium. Clin Chem Lab Med 2017 (in press) 2. Xie T*, Stathopoulou MG*, de Andrés F, Siest G, Murray H, Martin M, Cobaleda J, Delgado A, Lamont J, Peñas-LIedó E, LLerena A, Visvikis-Siest S. VEGF-related polymorphisms identified by GWAS and risk for major depression. Transl Psychiatry. 2017 Mar 7;7(3):e1055. *equal first authors 3. Chatelin J*, Xie T*, Kranjc T*, Stathopoulou MG, Toupance S, Rancier M, Vance D, Doherty L, Masson C, Murray H, Lamont J, Fitzgerald P, Weryha G, Benetos A, Marc J, Visvikis-Siest S. VEGF and osteoporosis; A VEGF Consortium collaborative study. The role of epistatic interactions between 4 VEGF-related SNPs on osteoporosis. Under preparation equal first authors 4. Xie T*, Stathopoulou MG*, Akbar S, Oster T, Siest G, Yen FT, Visvikis-Siest S. Effect of LSR polymorphism on blood lipid levels and age-specific epistatic interaction with the APOE common polymorphism. The effect of LSR functional polymorphisms on blood lipid traits and the role of the epistatic interactions between 2 candidate genes: APOE and LSR. Clin Genet (submitted). *equal first author 5. Xie T*, Akbar S*, Stathopoulou MG*, Oster T, Masson C, Yen FT, Visvikis-Siest S. The association of the epistatic interaction of APOE and LSR genetic variants in Alzheimer’s disease. The role of the epistatic interactions of APOE and LSR genetic variants on Alzheimer’s disease. Under submission

equal first authors. 6. Xie T*, Gorenjak V*, Stathopoulou M*, Dade S, Vance D, Doherty L, Marouli E, Masson C, Murray H, Lamont J, Fitzgerald P, Deloukas P, Visvikis-Siest S. Epigenome-wide association study in healthy individuals shows significant correlation with cg16170243 and waist circumference level. Epigenome-wide association study on central obesity phenotypes. Under submission *equal first authors 7. Xie T*, Gorenjak V*, Stathopoulou M*, Marouli E, Vance D, Doherty L, Masson C, Murray H, Lamont J, Fitzgerald P, Deloukas P, Visvikis-Siest S. EWAS of blood lipids in healthy population from STANISLAS Family Study (SFS). Epigenome-wide association study on blood lipids. Under submission *equal first authors

List of review publications

1. Chatelin J*, Stathopoulou MG*, Arguinano AA*, Xie T*, Visvikis-Siest S*. Pharmacogenomic Challenges in Cardiovascular Diseases: Examples of Drugs and Considerations for Future Integration in Clinical Practice. Curr Pharm Biotechnol. 2017;18(3):231-241. *equal contribution

Other publications

1. Aldasoro Arguinano AA, Dadé S, Stathopoulou M, Derive M, Coumba Ndiaye N, Xie T, Masson C, Gibot S, Visvikis-Siest S. TREM-1 SNP rs2234246 regulates TREM-1 protein and mRNA levels and is associated with plasma levels of L-selectin. PLoS One. 2017 Aug 3;12(8):e0182226.

2. Visvikis-Siest S, Aldasoro Arguinano AA, Stathopoulou M, Xie T, Petrelis A, Weryha G, Froguel P, Meier-Abt P, Meyer UA, Mlakar V, Ansari M, Papassotiropoulos A, Dedoussis G, Pan B, Bühlmann RP, Noyer-Weidner M, Dietrich PY, Van Schaik R, Innocenti F, März W, Bekris LM, Deloukas P. 8th Santorini Conference: Systems medicine and personalised health and therapy, Santorini, Greece, 3-5 October 2016. Drug Metab Pers Ther. 2017 May 24;32(2):119-127

Présentation synthétique de la thèse

Situation du sujet

Mon projet de thèse a été réalisé dans l'Unité de recherche UMR INSERM U1122 Interactions Gène-Environnement en Physiopathologie Cardio-Vasculaire (IGE-PCV). La recherche de l'Unité est axée sur les composantes génétiques et environnementales des phénotypes intermédiaires impliqués dans la pathophysiologie des maladies chroniques, en particulier des maladies cardiovasculaires (MCV).

Les anomalies de l'homéostasie lipidique sont des facteurs de risques importants, qui conduisent à différentes formes de dyslipidémies, d'hypertriglycéridémie et d'hypercholestérolémie. Il a été démontré que le dérèglement de l'homéostasie des lipides affecte divers troubles, notamment l'obésité, le diabète de type 2, les MCV, l'ostéoporose et les maladies neurodégénératives telles que la dépression majeure et la maladie d'Alzheimer (MA). Ces pathologies sont un lourd fardeau pour les patients, leurs familles et le système de santé. En revanche, le maintien de l'état normal des lipides peut réduire le risque de ces maladies. Cependant, le métabolisme des lipides est un système d'interaction complexe entre divers récepteurs des lipoprotéines, enzymes et transporteurs. Il est important de comprendre les mécanismes sous-jacents qui lient les lipides et les maladies chroniques et d'identifier l'existence de facteurs de risque communs entre les maladies.

L'angiogenèse est un autre mécanisme impliqué dans les maladies chroniques. Le facteur de croissance endothélial vasculaire (VEGF) est une molécule d'angiogenèse clé qui est associée à une variété de maladies humaines. Il est hautement héritable et il a une biologie complexe qui démontre son rôle important dans les processus physiologiques et pathologiques. Il représente un biomarqueur idéal qui pourrait servir de dénominateur commun des maladies chroniques.

Bien que des progrès significatifs aient été observés dans la génétique des maladies chroniques avec les approches classiques, telles que ‘gène candidat’ (gènes impliqués dans les voies métaboliques liées aux maladies chroniques) et GWAS (étude

d'association pangénomique), l'identification des variants génétiques communs pour les maladies chroniques est encore très limitée et peu des études ont examiné cette hypothèse. De plus, peu des variants fonctionnels ont été proposés en clinique. Ceci est dû principalement à une grande partie par le manque de réplication de certains résultats et de l'héritabilité des phénotypes intermédiaires inexpliquée, l'héritabilité dite manquante, qui pourraient pourtant, au moins en partie, s'expliquer par les interactions gène-gène (épistasie) et la méthylation de l'ADN (ADNm).

En se basant sur ces hypothèses, l’Unité U1122 a développé une stratégie intégrative basée sur des populations familiales saines et sur des études de phénotypes intermédiaires avec une approche transversale et / ou longitudinale, en particulier avec: L'étude STANISLAS Family Study (SFS), enquête longitudinale auprès de 1006 familles de Vandoeuvre-lès-Nancy, recrutées entre 1993 et 1995 (Visvikis-Siest et Siest 2008). Cette population ainsi que d'autres disponibles au Centre de Ressources Biologiques (BRC) IGE-PCV (BB-0033-00051) ont été utilisées dans cette thèse pour l'étude de l'effet de l'épistasie et de l'ADNm sur la variabilité des phénotypes intermédiaires des maladies chroniques.

Objectifs du travail

L'objectif principal de cette thèse était de mettre en évidence des associations entre trois biomarqueurs [VEGF, apolipoprotéine E (APOE), "lipolisis-stimulated receptor" (LSR)] impliqués dans quatre maladies chroniques (MCV, MA, ostéoporose et dépression) en utilisant diverses approches [gène-candidat, GWAS, épistasie et épigénétique (EWAS)] afin de trouver les liens entre les maladies et les mécanismes physiopathologiques communs impliqués. Ainsi, nous avons démontré :

1. L'importance des interactions épistasiques entre les polymorphismes liés au VEGF affectant les phénotypes intermédiaires cardiovasculaires dans un consortium international, le Consortium VEGF.

2. L’association du polymorphisme rs4416670, un polymorphisme identifié par GWAS lié aux taux de VEGF et aux facteurs de risque CVD, à un risque accru de dépression.

3. Une régulation génétique commune entre les marqueurs génétiques du VEGF et l'ostéoporose dans les populations post-ménopausées.

4. Une interaction fonctionnelle du polymorphisme LSR avec le polymorphisme commun de l'APOE inversant l'effet protecteur de l'allèle ε2 sur les taux de lipides d'une manière spécifique à l'âge.

5. Des interactions entre LSR et APOE qui pourraient prédire la susceptibilité d'un individu à la maladie d'Alzheimer.

6. Une association entre méthylation de l'ADN et tour de taille (WC) qui modifierait la transcription anti-sens de AC090241.2 affectant la transcription de ST8SIA5 et son activité enzymatique, par conséquence les gangliosides, suggérant ainsi la voie de signalisation d'insuline potentiellement impliquée.

7. Des associations entre méthylation de l'ADN et triglycérides (TG) capables d'affecter différentes maladies chez l'homme, révélant l'existence de liens complexes entre méthylation de l'ADN et maladies chroniques. Cela pourrait aboutir à de nouvelles possibilités thérapeutiques.

Résultats obtenus - discussion

Publication 1

Pharmacogenomic Challenges in Cardiovascular Diseases: Examples of Drugs and Considerations for Future Integration in Clinical Practice.

Défis pharmacogénomiques dans les maladies cardiovasculaires: exemples de médicaments et considérations pour une intégration future dans la pratique clinique.

Dans cette revue, nous présentons les défis, les progrès et les perspectives des médicaments contre les maladies cardiovasculaires et l'application de la pharmacogénomique dans la pratique médicale. L'efficacité et les effets indésirables des médicaments contre les maladies cardiovasculaires y sont également discutés. Les connaissances actuelles de ces médicaments associées aux connaissances du génome humain donnent l'espoir d'une révolution future dans le développement de médicaments qui optimisera le bénéfice des patients tout en réduisant le risque d'effets indésirables. La prise en compte des interactions gène × gène × environnement et l’introduction de données «omiques» dans les études pharmacogénomiques des médicaments contre les maladies cardiovasculaires faciliteront l’obtention de résultats fiables et favoriseront des traitements adaptés. De nouvelles stratégies de recherche pourront ainsi être envisagées ainsi que le développement de nouveaux médicaments.

Publication 2

A transnational collaborative network dedicated to the study and applications of the Vascular Endothelial Growth Factor–A in medical practice: The VEGF Consortium

Un réseau de collaboration transnational dédié à l'étude et aux applications du facteur de croissance endothélial vasculaire-A dans la pratique médicale : le consortium VEGF

Le VEGF European Genomic Federation - Consortium VEGF (www.vegfconsortium.org) a été fondé en juin 2014 par un groupe international de chercheurs s'intéressant au VEGF et à ses implications en médecine personnalisée.

Le consortium suit une stratégie de «biologie des systèmes» en intégrant une variété de méthodologies expérimentales, théoriques et computationnelles pour le développement d'outils «omiques» du VEGF cliniquement applicables en médecine personnalisée.

Dans le cadre du consortium, j'ai effectué une analyse des interactions épistatiques entre les polymorphismes liés au VEGF et les phénotypes intermédiaires des maladies cardiovasculaires dans la population SFS. Dans ce projet, 4 autres populations sont incluses dans les analyses. Une méta-analyse a également été réalisée et une réplication dans UK Biobank est en cours.

Dans la SFS, aucune des 15 combinaisons testés n'a atteint le niveau de significativité pour les phénotypes évalués. Néanmoins, dans la méta-analyse avec toutes les

populations, des interactions significatives ont été identifiées pour la pression artérielle et l’index d’adiposité viscéral (VAI).

En identifiant des interactions épistasiques significatives entre les polymorphismes mononucléotidiques (SNPs) liés au VEGF et les phénotypes intermédiaires des MCV, nous avons validé une partie des objectifs de travail du consortium VEGF, à travers une approche épidémiologique génétique des populations supposées saines. Cette étude est en cours car la réplication au sein du UK Biobank n’est pas encore terminé.

Publication 3

VEGF-related polymorphisms identified by GWAS and risk for major depression

Les polymorphismes liés au VEGF identifiés par GWAS et le risque de dépression majeure

Dans cette étude, nous avons évalué les associations de 4 polymorphismes liés au VEGF (rs10738760, rs6921438, rs6993770, rs4416670) avec le risque de dépression chez des patients (n = 437 sujets) et des sujets sains (n = 477 sujets). L'analyse a été réalisée en utilisant le logiciel GMDR0.9 (réduction dimensionnelle multifactorielle généralisée) qui est une alternative non-paramétrique pour détecter et caractériser les interactions non-linéaires entre les composantes génétiques et environnementales.

Parmi les quatre SNPs étudiés, seul le rs4416670 était significativement associé à un risque accru de dépression, tandis que l'allèle C mineur était plus fréquent chez les patients. Aucune interaction significative n'a été identifiée.

Dans ce projet, nous avons détecté une régulation génétique commune entre les marqueurs génétiques du VEGF et la dépression. L'identification de SNPs affectant le risque de développement de dépression pourrait être impliquée dans la dépression

résistante au traitement. L'identification du rôle fonctionnel de ces SNPs pourrait permettre la prédiction du risque de dépression majeure personnalisée et un traitement efficace.

Publication 4

The role of epistatic interactions between 4 VEGF-related SNPs on osteoporosis

Le rôle des interactions épistatiques entre 4 SNPS liés au VEGF sur l'ostéoporose

Le VEGF est une molécule hautement héréditaire impliquée dans de nombreuses maladies chroniques telles que les maladies osseuses. En outre, les cellules ostéoblastiques murines et humaines produisent le VEGF agissant dans la régulation paracrine de la formation des vaisseaux sanguins dans le tissu osseux et dans la régulation autocrine de l'ostéogenèse. D'après les résultats du consortium VEGF, 10 polymorphismes sont associés aux taux plasmatiques du VEGF. Nous avons supposé que ces SNPS pourraient avoir un lien avec les phénotypes de l'ostéoporose. Nous avons testé cette hypothèse sur les populations de femmes ménopausées. Le logiciel R a été utilisé pour les analyses.

Nous avons identifié l'interaction rs6921438 x rs1073876 qui affecterait la densité minérale osseuse (BMD) du col du fémur chez les plus jeunes femmes. La méta-analyse a identifié l'interaction rs6921438 x rs6993770 qui n'était pas évidente dans l'analyse des populations prises individuellement, probablement en raison de la puissance accrue de la méta-analyse. Dans une étude précédente, nous avons identifié que l'interaction rs6921438 x rs6993770 était associée à des niveaux d'interleukine 6 (IL6). Le lien entre ostéoporose et masse osseuse avec IL6 est bien établi, en particulier chez les femmes ménopausées (Zhao 2012). Par conséquent, une explication de l'effet de l'interaction épistasique entre les SNPs associés au VEGF

rs6921438 x rs6993770 sur la BMD du col du fémur pourrait être attribuable aux effets parallèles sur les taux d'IL6.

Publication 5

The effect of LSR functional polymorphisms on blood lipid traits and the role of the epistatic interactions between 2 candidate genes: APOE and LSR

L'effet des polymorphismes fonctionnels du LSR sur les traits lipidiques sanguins et le rôle des interactions épistasiques entre 2 gènes candidats: APOE et LSR

Le LSR est un récepteur des apolipoprotéines APOB et APOE qui participe à l'élimination des lipoprotéines riches en TG. L’APOE est un facteur de risque important pour les maladies cardiovasculaires, en particulier dans l’hyperlipoprotéinémie (HLP) type III. De plus, le gène LSR est situé en amont de l'APOE. Des adultes non-apparentés (n = 432) et des enfants non-apparentés (n = 328) ainsi que la cohorte ApoEurope ont été utilisés pour détecter l'implication du LSR dans les MCV par l'étude des interactions épistasiques du LSR avec l'APOE.

LSR rs916147 et APOE ε2 interagissent chez l'adulte pour les taux de cholestérol (TC), APOE et lipoprotéine de basse densité (LDL). Aucune interaction significative n'a été observée chez l’enfant. Nous avons également calculé l'effet direct du polymorphisme d'APOE sur les lipides sanguins. Ces données ont révélé que seul APOE ε2 chez les adultes était associé à des niveaux bas de TC, APOE et LDL, alors que son interaction épistasique avec rs916147 de LSR entraînait une augmentation des niveaux des traits lipidiques. Cela nous a amené à la conclusion que l'interaction épistasique du LSR SNP rs916147 avec APOE ε2 inversait l'effet direct de l'APOE ε2 sur les taux de TC,

d'APOE et de LDL. Dans la population de réplication, des effets similaires ont été observés pour les taux de TC et d'APOE.

Cette étude représente la première démontrant que LSR SNP rs916147 peut affecter les niveaux de lipides chez l'homme et révèle l'existence de liens moléculaires complexes entre LSR et APOE qui mériteraient des investigations plus approfondies pour identifier les mécanismes sous-jacents de HLP type III.

Publication 6

The role of the epistatic interactions of APOE and LSR genetic variants on Alzheimer’s disease.

Le rôle des interactions épistasiques entre les variants génétiques APOE et LSR sur la maladie d'Alzheimer.

Le polymorphisme commun de l'APOE (allèle ε4) est le facteur de risque le plus important pour la MA tardive. Le gène APOE humain est situé sur le chromosome 19q13.1, une région liée à la maladie d'Alzheimer. Un autre gène candidat localisé aussi dans cette région est le LSR. En tant que récepteur de l'APOE, LSR joue un rôle important dans l'élimination des lipoprotéines riches en triglycérides et régule donc l'homéostasie lipidique à la fois dans la périphérie et dans le système nerveux central. Une population européenne cas-témoins de la MA a été utilisée (142 patients atteints d'Alzheimer et 63 témoins) pour détecter l'association de LSR avec MA directement ou par interaction avec le polymorphisme commun de l'APOE.

La combinaison d'allèles non- ε4 (ε2 / 2, ε3 / 2, ε3 / 3) interagissait significativement avec les deux SNPs de LSR, indiquant que les porteurs des allèles non-ε4 avaient un risque accru de maladie. En raison du rôle du LSR comme récepteur de l'APOE dans l'homéostasie des lipides, sa localisation chromosomique (chromosome 19) ainsi que 25

des interactions APOE-LSR, cette molécule représente un candidat intéressant de prédiction de risque MA.

Publication 7

Epigenome-wide association study on central obesity phenotypes.

Étude d'association épigénétique avec les phénotypes de l'obésité centrale.

Les changements épigénétiques modulent l'expression génétique. Ils sont donc capables d'affecter un phénotype particulier et par conséquence les mécanismes physiologiques ainsi que la physiopathologie de nombreuses maladies dans lesquelles ils sont impliqués. Nous avons réalisé un EWAS sur les phénotypes d'obésité centrale chez des adultes et des enfants en bonne santé, à partir du premier recrutement de SFS (220 sujets de 75 familles) visant à étudier l'effet des modifications dues à la méthylation de l'ADN.

Les résultats ont montré une voie de signalisation de l'insuline qui pourrait expliquer l'association entre la méthylation du Cg16170243 et le tour de taille, par modification de la transcription anti-sens d’AC090241.2 affectant la transcription de ST8SIA5 et son activité enzymatique, influençant ainsi les gangliosides. Ces mécanismes doivent être vérifiés mais on sait que la dérégulation de signalisation d'insuline peut contribuer de manière significative à la résistance systémique à l'insuline dans l'obésité centrale.

Publication 8

Epigenome-wide association study on blood lipids.

Étude d'association épigénétique avec les lipides sanguins.

Les mécanismes épigénétiques pourraient être impliqués dans la régulation de la variabilité interindividuelle des niveaux de lipides et ainsi jouer un rôle important dans le profil de risque cardiovasculaire. Nous avons réalisé un EWAS chez des adultes et enfants en bonne santé à partir du premier recrutement de SFS (220 sujets de 75 familles) sur des lipides sanguins: lipoprotéine de haute densité, lipoprotéine de basse densité, triglycérides et cholestérol total.

Des associations significatives avec les taux de triglycérides (FDR p <0,05) ont été identifiées pour quatre sondes individuelles incluent TMEM150A sur le chromosome 2, PRKAG2 sur le chromosome 7, KREMEN2 (Kringle Containing Transmembrane Protein 2) sur le chromosome 16 et WDR83OS sur le chromosome 19. Les deux sondes près des gènes TMEM150A et WDR83OS ne sont pas encore associées à des maladies, mais semblent représenter deux nouveaux gènes candidats clès à étudier attentivement dans le futur.

Le site de méthylation de l'ADN cg08897188 est situé dans PRKAG2, responsable de la cardiomyopathie hypertrophique familiale (CHM). Ce site de méthylation est capable de lier les facteurs du doigt de zinc 1 [MZF1 associés à cardiomyopathie / dysplasie arythmogène du ventricule droit (ARVC/D)]. Le gène PRKAG2 pourrait être un candidat clé liant ARVD/C et CHM.

Le site CpG cg04580029 lié à KREMEN2 était significativement associé aux TG dans la population saine. Par conséquent, KREMEN2 pourrait affecter le risque pour MA et maladie coronarienne en bloquant la signalisation Wnt / bêta-caténine, dissociant / inhibant ainsi la formation du complexe LRP6-DKK1. KREMEN2 pourrait donc être proposé comme une nouvelle cible thérapeutique. Néanmoins, les mécanismes sous-jacents doivent être élucidés.

Actuellement, nos résultats représentent la première étude démontrant une association entre méthylation de l'ADN et TG qui pourrait affecter le risque pour différentes maladies chez l'homme et révéler l'existence de liens complexes entre méthylation de l'ADN et maladies chroniques majeures.

Conclusions et perspectives

Au cours des trois années de thèse, j'ai étudié des interactions épistasiques entre des polymorphismes identifiés par les approches comme GWAS et gène-candidat sur plusieurs maladies chroniques. L'épistasie peut être un outil puissant dans l'étude des maladies humaines.

Notre stratégie a révélé des interactions significatives avec tous les phénotypes évalués qui peuvent être différenciées selon l'âge.

Plus précisément, nos résultats suggèrent des interactions entre les polymorphismes des gènes LSR et APOE qui pourraient affecter le risque de HLP type III et de MA. Ainsi, nous pourrions proposer LSR comme modificateur de l'effet de l'APOE expliquant le lien paradoxe entre APOE et HLP type III et MA. La prochaine étape consiste à explorer et comprendre les mécanismes moléculaires sous-jacents de ces interactions.

De plus, les associations détectées entre méthylation de l'ADN, WC et TG suggèrent un nouvel aperçu des mécanismes moléculaires communs aux maladies chroniques susceptible de contribuer à des nouvelles possibilités thérapeutiques.

Les modifications épigénétiques identifiées pourraient altérer les voies moléculaires modulant l'expression d'un génotype en un phénotype particulier, ainsi affectant les mécanismes physiologiques et pathologiques impliqués dans les maladies cardiovasculaires, la dépression et l'ostéoporose, dans lesquelles les SNPs liés au VEGF sont également impliqués.

En effet, vue la grande diversité des fonctions du VEGF et les résultats de sa génétique obtenus précédemment ainsi que par cette thèse, le VEGF pourrait être proposé comme un biomarqueur potentiel servant comme régulateur commun, mais avec des profils moléculaires spécifiques, de développement des maladies cardiovasculaires, de dépression et des maladies osseuses.

Pris ensemble, nos résultats montrent l'interrelation entre maladies chroniques et variants génétiques de VEGF, LSR et APOE. Dans les étapes suivantes, tous les résultats de cette thèse seront explorés au niveau fonctionnel pour comprendre les mécanismes détaillés et les interactions moléculaires impliquées.

Foreword

My thesis project was realised in the research Unit UMR INSERM U1122 Gene-Environment Interactions in Cardio-Vascular Physiopathology (IGE-PCV). The Unit’s research is focused on the genetic and environmental components of intermediate phenotypes implicated in chronic diseases pathophysiology, especially cardiovascular diseases (CVD).

The abnormalities in lipid homeostasis are important risk factors, which lead to different forms of dyslipidemias, hypertriglyceridemia and hypercholesterolemia. Dysregulation of lipid homeostasis has been shown to affect various disorders, including obesity, type 2 diabetes (T2D), CVD, osteoporosis and neurodegenerative diseases like major depressive disorder (MDD) and Alzheimer’s disease (AD), all of which cause a high burden for patients, their families and the healthcare system. In contrary, the maintenance of normal lipid status can reduce the risk for those diseases. However, lipid metabolism is a complex interaction system between various lipoprotein receptors, enzymes and transporters. It is important to understand the underlying mechanisms that link lipids and chronic diseases and to identify the existence of common between diseases risk factors.

Angiogenesis is another mechanism that is involved in chronic diseases. The vascular endothelial growth factor (VEGF) is a key angiogenesis molecule that is associated with a variety of human disease. It is highly heritable and it has a complex biology that demonstrates its important role in the physiological and pathological procedures. It represents an ideal biomarker that could serve as a common denominator of chronic diseases.

Although significant progress has been observed in the genetics of chronic diseases with the classical approaches, ‘candidate gene’ (genes involved in metabolic pathways related to chronic diseases) and GWAS (genome-wide association study), the identification of genetics variants that represent common risk for chronic diseases is still limited. Moreover, little is achieved in clinical implementation of biomarkers identified. This is mainly due to lack of replication of results and to a large part of 31

heritability that remains unexplained, the so-called missing heritability, which could be partly explained by gene-gene (epistatic) interactions and DNA methylation (mDNA).

In view of the above points, the Unit U1122 developed an approach based on apparently healthy familial populations and on genotype-intermediate phenotypes investigations from a transversal and/or a longitudinal point of view, with in particular: The STANISLAS Family Study (SFS) is a 10-year longitudinal survey involving 1006 families from Vandoeuvre-lès-Nancy, France, recruited between 1993-1995 (Visvikis-Siest and Siest 2008). This population along with others available in the Biological Resources Centre (BRC) IGE-PCV (BB-0033-00051) was used in this thesis for the study of the effect of epistasis and mDNA on the variability of intermediate phenotypes for chronic diseases.

The principal goal in this thesis was to delineate the association between molecules and biomarkers that are involved in chronic diseases, using a variety of approaches, including ‘candidate gene’ and GWAS, and methods for the assessment of the role of epistasis and epigenetic modifications, in order to investigate the links between diseases and the involved pathophysiological mechanisms related to key molecules. In this study, the principal markers assessed include the VEGF, apolipoprotein E (APOE), lipolysis-stimulated lipoprotein receptor (LSR).

Introduction

Chronic diseases

Chronic diseases are long-term diseases that generally evolve slowly and affect all population. According to the World Health Organization, chronic diseases are major causes of mortality in the world, responsible for about 63% of mortality (World Health Organization 2016). Genetic components represent a great risk factor for many chronic diseases, due to the heritable component of diseases and diseases risk factors. Furthermore, the environment can also contribute to affect the risk for the diseases such as physical activity, diet, alcohol consumption, and smoking.

Age and gender are the primary risk factors for chronic diseases. The National Council on Aging indicated that about 80% of adults >65 years old have at least one chronic condition, while 68% have two or more as shown in Figure 1. Therefore, the prediction, diagnose and treatment of chronic diseases is a challenge.

Figure 1: 10 common chronic diseases for adults >65 years old. (https://www.ncoa.org/blog/10-common-chronic-diseases-prevention-tips/)

Cardiovascular diseases

Cardiovascular diseases (CVD) include coronary artery disease (CAD), cerebrovascular disease (hemorrhagic or ischemic, transient or constituted), peripheral arterial disease, rheumatic heart disease, and congenital heart disease. They constitute a major cause of death worldwide. According to the World Health Organization, about 17.5 million people died from CVD in 2012, with coronary heart disease (CHD) and stroke being the main reason (World Health Organization 2016).

Although the exact cause of CVD is still unclear, epidemiologic studies and randomised clinical trials have revealed that some disease risk factors could be prevented such as for CHD (Cooper, Cutler et al. 2000). In addition, the various underlying mechanisms of CVD are depending on the specific disease (GBD 2013 Mortality and Causes of Death Collaborators).The clinical phenotype of coronary artery and peripheral arterial disease is associated with atherosclerosis, for example. Important risk factors have been described such as age, gender, high blood pressure (hypertension), elevated levels of low-density lipoprotein cholesterol (LDL), smoking, and obesity (World Heart Federation , Jousilahti, Vartiainen et al. 1999, Stephen J. McPhee 2012). Thus, these diseases are clearly multifactorial, and the existence of a causal complicate relationship between diseases and factors exist.

The major CVD risk factors

There are many risk factors for CVD that generally, can be classified two parts, the modifiable and non-modifiable. Factors like age, gender, family history/genetic predisposition, are immutable, however, other risk factors could be changed or treated by environmental factor (lifestyle change, smoking cessation, social change, drug treatment).

Age and gender

Age is by far the immutable risk factor in developing CVD due to the aging of the body. Previous population studies have shown that age is associated with serum total cholesterol (TC) level (Jousilahti, Vartiainen et al. 1999). Compared to men who have increased levels of total cholesterol level around 45 to 50 years old, women are often affected more lately, generally around 60 to 65 years old (Jousilahti, Vartiainen et al. 1999). In addition, degeneration of arterial elasticity and compliance are associated with aging, leading to CHD (Jani and Rajkumar 2006), which can start even since childhood (D'Adamo, Guardamagna et al. 2015). Generally, the risk of heart disease in men is greater than pre-menopausal women. However, the risk in post-menopause women is similar to that of men (World Heart Federation), although this observation seems to be also affected by other factors such as country, region, age and level of economic development (Finegold, Asaria et al. 2013). Gender differences are a great risk contributor for CVD due to the changed hormonal states at different ages in women and men (Regitz-Zagrosek, Oertelt-Prigione et al. 2016). Estrogens are a major sex hormone that can explain the lower risk in pre-menopause women, as it is implicated in the glucose metabolism and hemostatic system and has also independent effects in the vascular endothelial cells (Mendelsohn and Karas 1999). In contrary, the risk of CVD is raised in post-menopause women, according to decrease of estrogens’ levels, which are related with changes in the lipid metabolism, that influence the atherogenesis by decreasing the high-density lipoprotein (HDL) cholesterol levels while increasing LDL (Grodstein , Stampfer et al. 1996).

Family inheritance

Family history is a well-known risk factor for CVD, which includes both shared genetics and environmental factors within a family. Adults with one or more first-degree relatives with CVD have greater risk that those with no relatives with CVD (Kardia, Modell et al. 2003). In addition, according to World Heart Federation 36

report, if both parents have CVD before 55 years old, the risk of developing CVD could rise to 50% for their offspring.

Tobacco

Cigarette smoking in both men and women is a major health hazard, with a great significance to CVD, although the underlying mechanisms are complex and to one extent unclear (Huxley and Woodward 2011). It is known that tobacco exposure increases oxidative stress that initiates dysfunction (Figure 2). In addition, cigarette smoking is considered to predispose to atherosclerotic syndromes and endothelial dysfunction (Centers for Disease, Prevention et al. 2010). Both active or passive (second-hand smoke) cigarette smoke exposure, predispose to cardiovascular events (Mendis, Puska et al. 2011).

Figure 2: The potential mechanism of cigarette smoking initiates cardiovascular dysfunction (Ambrose and Barua 2004).

Hypertension

Arterial hypertension is one of the most important independent risk factor for CVD. High blood pressure is generally defined as a systolic blood pressure ≥ 140 mmHg and/or a diastolic blood pressure ≥ 90 mmHg.

Hypertension is actually becoming the most common disease in the world. It leads to the hardening and thickening of both vessel walls and ventricular walls (O'Rourke, Safar et al. 2010) that lead to increased risk for CVD, including stroke, heart failure and CHD. Hence, according to clinical medicine research, reduction in blood pressure reduces the incidence of stroke (Rashid, Leonardi-Bee et al. 2003). Furthermore, high blood pressure is a common comorbidity in overweight population.

Obesity

Obesity is becoming a serious problem with epidemic proportions in both adults and children (Eckel, York et al. 2004). Currently, overweight and obesity are defined by body mass index (BMI: weight in kilograms/height2 in meters). In adults, overweight is defined as a BMI of 25.0 to 29.9 kg/m2 and obesity is defined as a BMI ≥ 30.0 kg/m2, which is associated with increased disease risk (Table 1).

Obesity is a negative impact on health for population that it is associated with numerous comorbidities such as CVD, type 2 diabetes, hypertension, and hyperlipidemia. In fact, obesity is an independent risk factor for CVD (Poirier and Eckel 2002), which has been reported to alter lipid metabolism profile in obese children (Teixeira, Sardinha et al. 2001). Additionally, elevated tumor necrosis factor-α (TNF-α), interleukin-6 (IL6), leptin, angiotensinogen, lipoprotein lipase, C-reactive protein (CRP) were related with higher BMI, and were observed in obese patients (Kern, Saghizadeh et al. 1995). Particularly IL-6 and CRP showed a strong correlation with obesity (Bastard, Jardel et al. 1999). Hence, the association between

obesity and the inflammatory system represents another underlying mechanism that can influence risk of CVD, even in the absence of other comorbidities.

Table 1: Classification of overweight and obesity by percentage of body fat, body mass index, waist circumference (WC) and associated diseases risk

Body Mass Index, Men ≤102 cm Men > 102 cm (kg/m²) Women ≤ 88 cm Women > 88 cm Underweight < 18.5 Normal 18.5 – 24.9 Overweight 25.0 – 29.9 Increased High Obesity, class I 30.0 – 34.9 High Very High II 35.0 – 39.9 Very High Very High III (extreme obesity) ≥ 40 Extremely high Extremely high Reference : (National Heart Lung and Blood Institute task force 1998)

Dyslipidemia and blood lipids

Dyslipidemia is defined as abnormal lipids’ levels in the blood, generally expressed as elevation of lipids’ levels, which are often affected by diet and lifestyle (Hendrani, Adesiyun et al. 2016). Previously study has shown that dyslipidemia is a major risk leading to CHD, it is mainly characterised by elevated total and LDL cholesterol, increased levels of triglycerides (TG), and sometimes decreased HDL cholesterol levels in the circulation (Sacks, Pfeffer et al. 1996), however, among them, the elevation of TG and a low level of HDL cholesterol are still a contentious issue (Libby 2015).

LDL and HDL cholesterol are two main transports that carry cholesterol to cells. Generally, HDL cholesterol is called the ‘good cholesterol’, which helps to keep the balance of cholesterol levels in the arteries, while LDL cholesterol is called the ‘bad cholesterol’, which is linked to accumulation of cholesterol in the arteries, thus leading to arterial wall thickening and heart disease (Solberg and Strong 1983). High blood cholesterol is therefore one of the major risk factors for CVD. LDL is the major cholesterol carrier lipoprotein in plasma, containing approximately 42% cholesterol esters (Davis and Wagganer 2005), and it is considered as a causal factor for CHD (Ference, Ginsberg et al. 2017) (Figure 3). Some genetic diseases have been linked with increased LDL levels. Familial hypercholesterolemia leads to elevated cholesterol concentration in plasma due to deficit of the LDL receptor (LDLR) due to mutations in the LDLR gene (Goldstein, Hobbs et al. 2001). The mutation in the apolipoprotein B (APOB) gene is another reason that increases the LDL levels in plasma due to decreased binding of ApoB-100 on each lipoprotein particle to LDL receptor (Nordestgaard, Chapman et al. 2013)..

HDL is the smallest of the lipoprotein particles, with apolipoprotein A1 (ApoA-I) being its major component (Colvin, Moriguchi et al. 1999). It is reported to be atheroprotective (Lüscher, Landmesser et al. 2014). Additionally, HDL is a major transporter, who takes cholesterol from body tissue to the liver, and about 30% cholesterol is transported by HDL in the blood (Toth 2005). Consequently, high HDL levels are reducing the risk of CVD (Choi, Vilahur et al. 2006). Hence, levels of HDL are negatively associated with risk of CVD. However, the causal relation between HDL and atherosclerosis is still unclear (Toth, Barter et al. 2013).

TG are the main constituents of body fat. They are also present in blood through lipoproteins such as chylomicrons and very low-density lipoprotein (VLDL) from and to the liver, which accomplish the bidirectional transference of adipose fat and blood glucose (Nabel 2003) (Figure 3). Furthermore, case-control studies have demonstrated that TG levels could be considered as an independent CVD risk factor (Brunner,

Altman et al. 1977, Austin, Breslow et al. 1988). Recent studies have provided evidence that high concentrations of TG in blood are strongly associated with low concentrations of HDL cholesterol, thus increasing the risk of CVD (Miller, Stone et al. 2011, Varbo, Benn et al. 2013), similar to high LDL.

Figure 3: Pathophysiology of atherosclerosis through cholesterol synthesis (Nabel 2003)

Major depressive disorder (MDD)

Depression is another chronic disease, which is a common, severe, disabling mental disease that affects millions of individuals of all ages, and it is among the leading causes of mortality worldwide. Furthermore, it is often a comorbidity of other chronic diseases (Noel, Williams et al. 2004). MDD is one of the most important depression disorders. The symptoms of MDD are expressed as very low mood that accompanies all aspects of life, never feeling any pleasure, combined with changes in cognition. MDD is a complex disease and it is difficult to establish stable diagnostic criteria, therefore, the diagnosis of MDD is based on the person's reported experiences and a mental status examination. The current diagnostic criteria of MDD depend on the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) of 2013 (American Psychiatric Association 2013).

Women are affected more often than men, although the cause is unclear and need to be elucidated (Kessler and Bromet 2013). Among all medical conditions, MDD is considered a great contributor (Vos, Barber et al. 2015), and is a risk factor of developing conditions related to chronic diseases such as diabetes mellitus, CVD (Whooley and Wong 2013), and Alzheimer's disease (AD) (Penninx, Milaneschi et al. 2013) (Figure 4). Furthermore, MDD is associated with a high rate of death by suicide, which is almost 20-fold more elevated compared to the general population (Chesney, Goodwin et al. 2014). It is a major burden to the patients, their families and the public health system. (Braun, Bschor et al. 2016).

Figure 4: MDD influences development of various diseases.

The risk factors for MDD

Genetic, molecular, and neuroimaging studies contribute to advances in understanding of MDD. These studies provide strong evidences that MDD has a high heritability, while GWAS have identify some loci (Hyde, Nagle et al. 2016) that are believed to be major risk factors. In addition, the environmental, psychological, major lifestyle changes, chronic health problems, substance abuse, and emotional abuse during childhood are also considered risk factor, which strongly increased risk for MDD (Li, D'Arcy et al. 2016). However, the mechanisms that explain the effect of the environment on genetics are still unclear. Although MDD is a complex disease, many studies focused on functional gene analysis and biological mechanisms have successfully identified some candidate genes with biological mechanisms and metabolic pathways that are associated with MDD and response to antidepressant (Serretti, Kato et al. 2007, Licinio, Dong et al. 2009, Kato and Serretti 2010). Although the cause of MDD is unknown, however, according to current knowledge of neurobiology of MDD (Figure 5), the changed hippocampal volumes and neural networks are influencing the risk for MDD (Etkin, Buchel et al. 2015). VEGF gene

has effects in the central nervous system (CNS) such as hippocampus (Yasuhara, Shingo et al. 2004). Thus, VEGF is considered as a potential biomarker for MDD.

Figure 5: Neural systems of relevance to major depressive disorder. DLPFC=dorsolateral prefrontal cortex. mPFC=medial prefrontal cortex. ACC=anterior cingulate cortex. OFC=orbitofrontal cortex. VLPFC=ventrolateral prefrontal cortex (Kupfer, Frank et al. 2012).

Osteoporosis

Osteoporosis is a major problem of public health. It is estimated that 200 million people suffer from osteoporosis in the world and 30 % women with menopausal status in United States of America and Europe present osteoporosis (NIH 2015). Osteoporosis is

defined as low bone mass and micro-architectural deterioration of bone tissue that are associated with a higher risk for increasing the fracture risk (Johnell and Kanis 2005). Postmenopausal women are greater contributors of the osteoporosis cases, due to the bone loss, which is related to reduce estrogen levels after menopause.

Bone mineral density (BMD) is a key criterion for the identification of osteoporosis in clinical medicine. T-score, which is the standard deviation (SD) of the BMD of an individual compared with the peak BMD of 30 years old adults, is the criterion for the definition of osteoporosis (Kanis 1994). However, these osteoporosis diagnostic criteria are not including men younger than 50 and premenopausal women (Siris, Adler et al. 2014).

In osteoporosis, increased osteoclastic bone resorption relative to osteoblastic bone formation results in low bone mass and fragility. At each step of skeleton development (embryonic stage, post natal growth) and bone remodeling, vasculature is important, bringing precursors of osteoclasts and osteoblasts, nutrients, vitamin D, growth and differentiation factors (Trueta J 1963, Brandi ML 2006). Vasculogenesis in involved in osteogenesis (Grellier M 2009) and angiogenesis plays a major role in skeletal development and fracture repair.

Alzheimer’s disease (AD)

AD is a complex multifactorial neurodegenerative disorder, which is the most common form of dementia in the elderly (Scheltens, Blennow et al. 2016). Brain contains approximately 20% of total body cholesterol, which is crucial in maintaining the cognitive functions (Zhang and Liu 2015, Chang, Yamauchi et al. 2017). Being an essential neuronal membrane component, cholesterol contributes to AD pathogenesis by altering the proteolytic processing of amyloid precursor protein via the membrane-bound cleavage enzymes that leads to low or high extracellular deposits of amyloid-beta peptide (Aβ), a major hallmark of AD (Grziwa, Grimm et al. 2003,

Xiong, Callaghan et al. 2008). Despite the presence of blood-brain barrier that precludes the exchange between cerebral and peripheral cholesterol pools, various epidemiological studies have reported the association of high plasma total cholesterol levels with an increased AD risk (Notkola, Sulkava et al. 1998, Kivipelto, Helkala et al. 2002, Solomon, Kivipelto et al. 2009). Moreover, any disturbances in lipid homeostasis, manifested in the form of dyslipidemia, have been shown to be causally linked to the development and progression of AD (Reitz 2013). In brain, the transport of cholesterol and other lipids from glial cells to neurons is mediated by apolipoprotein E (APOE) gene (Rebeck, LaDu et al. 2006) via its interaction with various lipoprotein receptors (Stenger, Pincon et al. 2012, Huang and Mahley 2014), while the APOE ε4 allele is the strongest risk factor for the late-onset AD (Kim, Basak et al. 2009).

Common biomarkers

Biomarkers are powerful tools in the war against chronic diseases. They are defined as “a characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention” by an national health institute working group (Biomarkers Definitions Working 2001). They can indicate a variety of health or disease characteristics, which are caused by environmental or genetic factors. They can be markers of subclinical or clinical disease, or indicators of the response to interventions. Biomarkers can therefore be used in many ways such as for early prevention, prediction and treatment of common chronic diseases.

Several biomarkers have been established in CVD, including APOE (Corder, Saunders et al. 1993, Poole, Walker et al. 1998), VEGF (Hutchinson 2013), blood lipids (Bachorik and Ross 1995, Stein and Myers 1995, Warnick and Wood 1995). Among them, APOE and VEGF are also considered as potential biomarkers for AD,

MDD and osteoporosis (Mayer, Bertram et al. 2005, Kim, Basak et al. 2009, Clark-Raymond, Meresh et al. 2014).

Comorbidities

Comorbidities are defined as any two or more diseases that occur in one person at the same time. Generally, more diseases can occur at the same time because of the existence of potential relationships between them. For example, one risk factor can be implicated in many diseases, and some diseases may also be risk factors for other diseases, such as obesity that is a risk factor for diseases like diabetes and CVD (Australian Institute of Health and Welfare 2016).

The figure 6 presents data related to comorbidities in patients with common chronic diseases such as chronic obstructive pulmonary disease (COPD), T2D, cancer, arthritis, CVD, back pain, mental health conditions, and asthma (Figure 6).

Figure 6: Comorbidity among CVD, arthritis, back pain and mental health conditions (Australian Institute of Health and Welfare 2016).

Furthermore, aging has a strong association with comorbidities. This is because along with aging, and decline of the immune system, the older people are more likely to develop many diseases and comorbidities (Figure 7).

Figure 7: Comorbidities among selected chronic diseases by age (Australian Institute of Health and Welfare 2016).

Epistasis

Epistasis is defined generally as the interaction between different genes that can affect phenotypes. In other words, in the presence of an interaction between two genes, one epistatic gene could influence the expression of another gene (Hurst 2000). Furthermore, human genetic disease is an area where the epistasis appears to be fairly common (Nagel 2005). Indeed, many studies have revealed that epistasis is involved in many complex diseases, such as diabetes (Moore 2003, Wiltshire, Bell et al. 2006) and CVD (Thornton-Wells, Moore et al. 2004). Evidence show that epistasis is also involved in other diseases, including bone diseases, neurological disorders like AD and depression (Koller, Liu et al. 2008, Combarros, Cortina-Borja et al. 2009).

Epigenetics

In 1942, 'epigenetics' were defined as changes in phenotype without the changes in genotype (Waddington 2012). Now days, the epigenetics changes are described as the study of changes in gene function that are mitotically and/or meiotically heritable and that do not entail a change in DNA sequence (Wu and Morris 2001), including mDNA, histone modifications, and chromatin remodeling. mDNA is the methylation of the DNA molecule without altering of the DNA sequence, which can change the activity of DNA. In mammals, mDNA is a post-replication modification mostly of cytosine at C5 in CpG dinucleotide (Bird 2002) (Figure. 8). CpGs are regions of DNA where a cytosine and a guanine are separated by one phosphate (Stevens, Cheng et al. 2013) and are very common in the human genome (28 millions). Among them, approximately 60-80% are methylated. Least 10% of CpGs occur in GC-rich regions called CpG islands. Moreover, if CpG islands in promoters are close to the transcription start sites, they could regulate gene expression by regulating chromatin structure and transcription factor binding (Deaton and Bird 2011). In mammalian cells, the DNA methyltransferase 1, 3A and 3B (DNMT1, DNMT3A and DNMT3B), which are responsible for maintaining mDNA, are essential for mammalian development (Okano, Bell et al. 1999). Many studies showed that mDNA is an important factor implicated in complex human disease, including cancer, cardiology, osteoporosis and neurodegeneration (Neidhart 2016).

Figure 8: DNA methylation, the methylation of cytosine from 5-methylcytosine (Qiu 2006)

Other modifications include the modification of histones, which show a diversity and complexity (Fischle, Wang et al. 2003). Histones package is a part of nucleosomes, which is around DNA and is playing a role in gene regulation. Histone modifications are covalent post-translational modifications to histone proteins. Depending on the changes of amino acids in histones, the modification of histones can regulate gene expression though modifying chromatin structure or recruiting histone modifiers, including methylation, phosphorylation, acetylation, ubiquitylation, and sumoylation (Kouzarides 2007) (Figure 9). The most commonly modified histones are H3 and H4. Histone modifications and DNA modifications could influence each other through direct interactions between histone and DNA methyltransferases during mammalian development that ameliorates our understanding of normal development as well as

development of diseases (Cedar and Bergman 2009). Indeed, links have been demonstrated in cancer through a process of 'selection' (Jones and Baylin 2007).

Figure 9: The most common histone modifications (abcam).

The targeted candidate biomarkers

Three candidate molecules were proposed by our laboratory in the study of common denominators of chronic diseases: VEGF, APOE, and LSR. With the use of candidate gene approaches, GWAS and DNA methylation related to these molecules we are trying to explain the “missing heritability” observed for the chronic diseases

intermediate phenotypes and to provide evidences for the interactions between common molecules involved in the pathophysiological mechanisms of chronic diseases.

Vascular endothelial growth factor (VEGF)

VEGF was first identified as a growth factor and a permeability factor of endothelial cells. It is mainly involved in angiogenesis, the formation of new blood vessels from pre-existing vessels in physiological situations such as pregnancy, wound healing (Yancopoulos, Davis et al. 2000), but it is also implicated in many pathologies such as diabetic retinopathy, rheumatoid arthritis, CVD and cancer (Ferrara, Gerber et al. 2003).

VEGF belongs to a family of vascular and lymphatic endothelial cell growth factors that includes the VEGF A (more commonly VEGF), B, C, D and placental growth factor (PIGF) (Table 2). In addition, it also exists VEGFE in parapoxvirus (Takahashi and Shibuya 2005) and snake venom VEGF (Suto, Yamazaki et al. 2005). VEGF exerts its action through binding with 3 VEGF tyrosine kinase receptors known as VEGF receptors 1, 2 and 3 (VEGFR1, VEGFR2, and VEGFR3) as well as co-receptor neuropilin and heparin sulphate proteoglycans (HSPGs) (Figure 10) (Ferrara, Gerber et al. 2003). The interaction of the members of VEGF and their receptors in different cell types could stimulate different signaling pathways such as the RAS / RAF / MEK / ERK and PI3 kinase / AKT / mTOR pathways that induces cell proliferation, migration and survival. The VEGF / VEGFR pathways are complex due to the different VEGF isoforms that may be present depending on the type of cell and the physiological / pathological procedures (Bates, Cui et al. 2002).

Table 2: Size of mRNA and protein of VEGF family members

Genes Chromosomal Size of mRNA (kb) Size of proteins (kDa) Location VEGF-A 6p23.1 3.7, 4.5 21 VEGF-B 11q13 1.4 21, 30 VEGF-C 4q34 2.4 20-21 VEGF-D Xp22.31 2.2 20-21 PIGF 14q24 1.7, 1.2 38, 30

Figure 10: VEGF receptor-binding properties and co-receptors (a) VEGF bind to the three VEGF receptor (VEGFR) tyrosine kinases. VEGFC and D could bind to VEGFR2 after proteolytic processing. (b) Co-receptor heparan sulphate proteoglycans (HSPGs) and neuropilins modulates VEGFR signaling. (Olsson, Dimberg et al. 2006)

VEGF has been identified by Senger et al. It is the best characterised and most studied member of the VEGF family. The VEGF-A gene gives rise to more than 12 isoforms grouped in two major families due to differential pre-mRNA splicing. The VEGF family results from splicing in exons 6, 7 and the proximal splice site in exon 8 (exon 8a), while the VEGFb family isoforms differ in the selection of a distal splice site at exon 8 (exon 8b) located 66bp downstream of the proximal splice site. The VEGF isoforms promote angiogenesis whereas the VEGFb isoforms are anti-angiogenic (Woolard, Bevan et al. 2009). In normal conditions in adults, the anti-angiogenic isoforms are expressed in non-angiogenic tissues and represent >50% of total VEGF protein. In contrast, many cancers are characterised by a switch from the anti-angiogenic to pro-angiogenic isoforms which leads to neo-vascularisation and consequently to tumor growth. Also, it has been suggested that the isoform profile of different tumors may affect the responsiveness to anti-VEGF therapies (Nowak, Amin et al. 2010). According to the characteristic of specific isoforms, VEGF can be linked with the development of human coronary atherosclerosis (Inoue, Itoh et al. 1998). The most common isoforms are the VEGF-121, VEGF-145, VEGF-165, VEGF-189 and VEGF-206 (Figure 11) (Tischer, Mitchell et al. 1991). Different biological effects of the VEGF isoforms can occur by their interaction with the cell surface receptors.

In addition, VEGF has been shown to affect the central nervous system in brain such as hippocampus (Cao, Jiao et al. 2004, Yasuhara, Shingo et al. 2004), and neurogenesis has been shown to be stimulated in vitro and in vivo by VEGF (Jin, Zhu et al. 2002). VEGF is recently considered as a potential biomarker for MDD, However, several studies have investigated the peripheral/serum levels of VEGF in depressed patients compared with healthy controls with unclear and ambiguous results (Clark-Raymond and Halaris 2013), due to differences in the studied populations’ characteristics (Isung, Mobarrez et al. 2012, Kotan, Sarandol et al. 2012, Lee and Kim 2012). However, Viikki et al have shown that the VEGF polymorphism rs699947 is associated with treatment resistant depression (Viikki, Anttila et al. 2010). A link between VEGF and depression could be the pro-inflammation cytokines, which are 56

implicated in depression (Howren, Lamkin et al. 2009). In addition, VEGF is regulated by pro-inflammation cytokines in depression (Schmidt, Shelton et al. 2011). Therefore, interactions between VEGF and pro-inflammation cytokines could influence depression and/or antidepressants efficacy (Audet and Anisman 2013).

VEGF is also implicated in bone diseases. Presence of VEGF receptors has been detected on primary osteoblasts (Midy V 1994). It has been also demonstrated that VEGF promotes bone remodeling acting directly on survival, chemotactic migration (Mayr-Wohlfart U 2002) and activity (Tan YY 2010) of osteoblasts. It has also been demonstrated that murine and human osteoblastic cells produce VEGF acting in the paracrine regulation of blood vessels formation in bone tissue and in the autocrine regulation of osteogenesis (Mayer H 2005, Street J 2009). A well-established protocol with engineered mice that they do not express VEGF in bone progenitor cells mimics the clinical phenotypes of osteoporosis with decrease in trabecular bone, thinner bone cortices and an increase in adipocytes associated with trabecular bone (Liu Y 2012).

Figure 11: Structure of the main splicing variants of VEGFA in humans (Olsson, Dimberg et al. 2006).

Apolipoprotein E (ApoE)

It is well known that the abnormalities in lipid homeostasis are important factors that lead to different forms of dyslipidemias, hypertriglyceridemia and hypercholesterolemia. Plasma cholesterol and TG are major risk factors for CVD (Hokanson and Austin 1996). Apolipoproteins play also an important role in the regulation of the processing and clearance of lipoproteins. Among them, apolipoprotein E (APOE) has been considered as an important risk factor for CVD (Upadhyay 2015). Furthermore, APOE, as ligand, is recognised by lipoprotein receptors in the periphery and it is the major transporter that facilitates cholesterol transfer from glial cells to neurons in the CNS (Mahley 2016, Liao, Yoon et al. 2017). In human APOE gene, there is a common polymorphism with three alleles - ε2, ε3 and ε4, which significantly affect the levels and variability of cholesterol metabolism (Boerwinkle, Visvikis et al. 1987). Among them, the allele ε3 is the most frequent allele and is considered as the neutral allele. As well know, APOE ε2/2 genotype of APOE gene is associated to type III hyperlipidemia (type III HLP; OMIM 107741). However, only few APOE ε2/2 individuals develop the disease (Evans, Beil et al. 2013). The allele ε4 has been implicated in atherosclerosis, and it is also the most important genetic risk factor for AD. It is shown that it stimulates amyloid precursor protein (APP) transcription and amyloid beta (Aβ) production (Huang, Zhou et al. 2017) and ischemic cerebrovascular disease (Lucotte, David et al. 1993, Siest, Bertrand et al. 2000).

Lipolysis-stimulated lipoprotein receptor (LSR)

The LSR gene encodes the lipolysis-stimulated lipoprotein receptor (LSR) that removes TG-rich lipoproteins during the postprandial phase (Yen, Roitel et al. 2008, Stenger, Corbier et al. 2012). Cell culture studies have demonstrated the ability of APOE-containing lipid particles to bind LSR (Yen, Mann et al. 1994) . Thus, LSR plays an important role in the regulation of lipoprotein uptake and metabolism

(Narvekar et al, 2009) and is linked to APOE. Indeed, LSR+/- mice revealed cholesterol homeostasis levels changes and also increased amyloid stress (Stenger, Pincon et al. 2012). Therefore, LSR regulates lipid homeostasis not only in the periphery, but also in CNS. The human LSR gene is located on chromosome 19q13. It is just upstream of the APOE locus, thus it is likely that there is a potential epistasis between these 2 genes implicated in lipids metabolism or AD pathologies. Blom et al. indicated that the chromosome 19q13.1 is associated with AD (Blom, Holmans et al. 2008). Two polymorphisms of LSR are very interesting due to their functionality. The rs916147 is an intron 5/ exon 6 variation of a splice junction that has been associated with plasma cholesterol, LDL and TG in obese adults (Brulliard, Ogier et al. 2008). In addition, this SNP can modify LSR expression in different tissues, thus indicating its potential functionality (http://www.gtexportal.org). The rs34259399 is a non-synonymous variation (S [Ser] N [Asn], 344 amino acid located between the regions coding for the transmembrane⇒ domain and the ligand binding site. Therefore, LSR could be a candidate biomarker in the study of AD.

Personalised medicine

Personalised medicine involves tailoring a health care for the patients, according to the personal history, genetic profile and specific biomarkers (Evans and Relling 2004). With the development of the Human Genome Project, personalised medicine using genetic information has been more successful in some fields such as oncology (King, Marks et al. 2003), and CVD (O'Donnell and Nabel 2008). Therefore, if our candidate molecules could reveal potential inter-relationships between the diseases and the underlying pathways with critical roles in the development of chronic diseases, we may be able to propose the use of these molecules as key factors for the implementation of personalised medicine in clinical practice.

Pharmacogenomics

Pharmacogenomics (PGx) is the complex study of the role of genome in drug response. It studies the effect of alterations in DNA sequence, chromosomal aberrations, and changes of the chromatin on drug response and effectiveness, as well as on side-effects.

Although there is still an ongoing debate about the cost of PGx, Dorfman et al. have indicated that PGx treatment optimization is feasible and can have a great value for many complex disease symptoms, because many medications can be used for the treatment of many comorbidities (Dorfman, Khayat et al. 2013). PGx is thus a feasible application for personalised medicine in chronic disease.

Pharmacogenomics can be applied in clinical research involving individual patient variations in the cytochrome P (CYP) 450 isoenzymes which are key enzymes in cancer formation and cancer treatment (Rodriguez-Antona and Ingelman-Sundberg 2006) and also anti-VEGF treatments (Meadows and Hurwitz 2012). Furthermore, PGx of several pharmacologic agents proposed for the treatment of dyslipidemia is a challenging field (Durrington 2003, Hachem and Mooradian 2006). The understanding of PGx can lead to achievements towards the goals for personalised medicine (Chatelin, Stathopoulou et al. 2017).

Hypothesis and objectives

As already described, the interrelations between chronic diseases are complex, due to the genetic variants and the environmental risk factors, which are essential components for the development of chronic diseases (WHO 2014). Furthermore, in patients with chronic diseases, usually the comorbidities could be often found in one patient at the same time, such as COPD, T2D, cancer, arthritis, CVD, bone diseases, mental health conditions, and others. Therefore, the biggest challenge of chronic diseases is to find the potential interrelationship between them and common risk factors using information on genetic variants and environment, which have the potential to improve or discover new strategies in personalised medicine for more efficient prediction, prevention and therapy.

In view of the technological revolution in the “-omics” era, numerous variants have been identified with a promising role in personalised medicine, thereby the individualisation is becoming more and more feasible. Nevertheless, the inter-individual variability of the heritable disease phenotypes is still poorly explained. This problem that is described as ‘‘missing heritability’’ could actually be explained by epistasic interactions. Indeed, the influence of SNPs is not merely additive, or dominant / recessive, but also interactive. In addition, gene-environment interactions may also play an important role. Similarly, epigenetic processes, including mDNA, may be equally important.

Thus, we hypothesised that the genetic biomarkers in multiple pathways may play a key role for the development of chronic diseases through epistasis and mDNA. These mechanisms could also generate molecular pathways possibly involved in chronic diseases and common risk factors. These results could then provide interesting material for the implementation of personalised medicine in clinical practice.

Among the selected molecules that were assessed in this thesis, APOE and VEGF have been extensively studied in the U1122 Unit. The Unit has revealed that APOE concentration represent an additional risk factor for AD, complementary and independent of the e4 allele (Siest, Bertrand et al. 2000). Concerning VEGF, the Unit 62

U1122 has previously identified 4 polymorphisms that explain almost 50% of the inter-individual variability of VEGF circulating levels through a GWAS (Debette, Visvikis-Siest et al. 2011). These polymorphisms were also associated with LDL and HDL levels (Stathopoulou, Bonnefond et al. 2013), and with adhesion and inflammation molecules in a healthy population (Azimi-Nezhad, Stathopoulou et al. 2013). Therefore, VEGF is considered as a key molecule that may offer applicable information with a very promising role in the implementation of personalised medicine in several common chronic diseases.

In this thesis, we have assessed the role of genetic variants associated with three molecules (APOE, LSR, VEGF) on the determination of individual variability of intermediate phenotypes of different chronic diseases (CVD, AD, depression and osteoporosis). Our strategy was to assess the effect of epistasis and epigenetics as these are considered important phenomena that can explain the “missing heritability”, observed for the chronic diseases intermediate phenotypes and can provide evidences for the interactions between common molecules involved in the pathophysiological mechanisms of chronic diseases. The goal of these hypotheses is to identify biomarkers that can be applied in risk prediction and management of diseases through personalised medicine.

The following projects have been developed:

1. The VEGF Consortium

2. The role of the VEGF-related SNPs on depression in a case-control study.

3. The role of epistatic interactions between all 4 VEGF-related SNPs on osteoporosis.

4. The effect of LSR functional polymorphisms on blood lipid traits and the role of the epistatic interactions with 2 candidate-genes: APOE and LSR.

5. The role of the epistatic interactions of APOE and LSR genetic variants in Alzheimer’s disease.

6. Epigenome-wide association study on central obesity phenotypes.

7. Epigenome-wide association study on blood lipids.

The specific objectives of the above projects that investigate the direct effects, epistatic interactions of SNPs and epigenetic modifications using candidate-genes, GWAS approaches, and DNA methylation are:

1. To develop a Consortium specific to the study of VEGF molecular determinants and their role in diseases stratifications. Within the frames of this consortium, epistatic interactions between the GWAS-identified SNPs related to VEGF were assessed against chronic diseases intermediate phenotypes.

2. To investigate the four GWAS identified SNPs related to VEGF in individuals with depression vs. healthy controls, in order to detect the associations between genetic biomarkers of VEGF and depression.

3. To investigate the epistatic effects of the GWAS identified SNPs related to VEGF with bone mineral density and osteoporosis in post-menopausal women.

4. To investigate the role of common functional LSR polymorphisms on blood lipids levels as well as the effect of potential epistasis of these SNPs with the APOE gene common polymorphism in childhood and adulthood in healthy populations.

5. To investigate the associations of the functional polymorphisms of LSR gene and its interactions with the APOE common polymorphism with risk for Alzheimer’s disease in a case-control study of patients with Alzheimer’s disease vs. healthy controls.

6. To investigate the association between DNA methylation CpG sites and central obesity phenotypes in SFS population, thus to discover underlying mechanisms linking the DNA methylation with central obesity related phenotypes.

7. To investigate the links between DNA methylation CpG sites and blood lipids in healthy populations, thus to discover underlying mechanisms linking the DNA methylation with blood lipids metabolism.

Materials and Methods

Populations

STANISLAS Family Study

The STANISLAS Family Study (SFS) is a 10-year longitudinal study that includes 1006 families residing in Vandoeuvre-lès-Nancy (France), recruited between 1993 and 1995 (Visvikis-Siest and Siest 2008). All subjects were of European-Caucasian origin. Exclusion criteria included the presence of chronic disorders (such as CVD or cancer) and an individual history of CVD, thus the individuals are considered as supposed healthy. The study protocol was approved by the local ethics committee and all participants gave their

informed written consent. The first recruitment (t0) included 1006 two-parent families (4455) of French origin with at least two children over six years of age. In the second visit, 75% of the families initially recruited were presented

(1998-2000, t5) and 375 families participated in the third visit (2003-2005, t10). This population is available at the BRC IGE-PCV “Gene-environment interactions in cardiovascular pathophysiology” - BB-0033-00051.

Cilento study

This is a population-based study that aims at identifying genetic risk factors for common diseases and traits. The sample includes isolated populations from three villages (Campora, Gioi and Cardile). The overall sample size of individuals participating to the study is 2100. All inhabitants of the selected isolated populations are invited to participate to the study. The decision to participate is voluntary and all participants signed an informed consent (Colonna, Nutile et al. 2007, Colonna, Nutile et al. 2009).

LifeLines Cohort Study and Biobank (LLs)

The general aim of the LLs Cohort Study is to unravel how life-time exposure to (universal) risk factors influences individual susceptibility to multifactorial diseases. LLs is a multi-disciplinary prospective population-based cohort study examining the health and health-related behaviours of 167,000 persons living in the North East region of The Netherlands in a three-generation design. It employs a broad range of investigative procedures in assessing the biomedical, socio-demographic, behavioural, physical and psychological factors which contribute to the health and disease of the general population, with a special focus on multi-morbidity and complex genetics. All survey participants were between 18 and 90 years old at the time of the enrollment (Stolk, Rosmalen et al. 2008, Klijs, Scholtens et al. 2015, Scholtens, Smidt et al. 2015, Nolte and van der Most 2017).

ApoEurope cohort

The ApoEurope cohort consists of 3706 men and 3228 women, aged 25–64 years, of European origin: Crete (Greece), Helsinki (Finland), Belfast (Northern Ireland), Nancy (France), Lisbon (Portugal), and Barcelona (Spain). This population is also available at the BRC IGE-PCV. This project was conducted with subjects not taking antihypertensive medications. Collection of samples and data and biological measurements were carried out by each recruitment center with standardised protocols and methods, which were similar to those applied in the SFS population (Schiele, De Bacquer et al. 2000).

Hypertensive population

A population of hypertensive individuals, who are not diagnosed with any CVD or other chronic diseases and are not treated with anti-hypertensive and other CVD treatments is available at the BRC IGE-PCV (n=945). Collection of samples and data and biological measurements were carried out using standardised protocols and methods, which were similar to those applied in the SFS population.

Case-control study of depression

Four hundred thirty seven unrelated Spanish Caucasians patients diagnosed with depressive disorders (Males: 132, Females: 305) who were attending a Mental Health Centre were included. The diagnosis of depressive disorders symptoms was based on the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV criteria) (APA 1994). A group control of 477 (Males: 187, Females: 290) Caucasians healthy volunteers from the same geographical region was recruited as the reference. The individuals of the control group were strict selected by psychiatrists who performed a mental and physical revision First-degree relatives were also assessed by psychiatrists regarding potential psychiatric disorders. However, to avoid familial stratification issues, only unrelated individuals have been included in the control group. Healthy volunteers with a previous history of adverse drug effects and those with any drug intake in the 2 weeks before the study were excluded from participation in the study. Women who reported or suspected pregnancy were also excluded from the study.

Populations of postmenopausal women:

The Slovenian cohorts consist of two different populations: 52 unrelated younger post-menopausal women, and 91 unrelated older osteoporotic post-menopausal women. The third population included 97 unrelated post-menopausal women from ARTEOS study, which was designed to investigate the relationship between osteoporosis and arterial rigidity within fractured patients’ cohort (https://clinicaltrials.gov/ct2/show/NCT01587456).

Alzheimer’s case-control population

A total of 205 unrelated individuals of European origin (Spain, Ireland, and Former Yugoslav Republic of Macedonia) participated in the study. Among them, 142 patients were diagnosed with AD. A control group of 63 healthy volunteers from the same geographical region was also studied. The subjects with the diseases, such as Huntington’s, Pick’s, Wilson’s, Creutzfeld-Jacobs’, or Parkinson’s diseases, normal pressure hydrocephalus, cerebrovascular, cardiovascular, or coronary heart diseases, inflammation or cancer, were excluded from the study. The project was approved by the related local ethics committees and all subjects gave their informed consent to participate in the study (Siest, Bertrand et al. 2000).

Genotyping – genome-scale methylation assay

Blood samples were taken between 8 and 9 a.m. after overnight fasting. Whole blood DNA was extracted by the Miller technique (Miller, Dykes et al. 1988) and was stored at – 80 °C until further use. DNA has been extracted from all participants and relative biobanks have been constructed in the BRC IGE-PCV.

The GWAS identified VEGF related SNPs rs6921438, rs4416670/rs1740073, rs6993770, rs10738760/rs2375981, rs2639990, rs4782371, rs10761741, rs7043199, rs34528081, rs114694170 were genotyped using the competitive allele-specific PCR chemistry coupled with a FRET-based genotyping system (http://www.lgcgroup.com/services/geno), and genotyping was confirmed by Randox Laboratories (Crumlin, UK; Evidence Investigator) using an assay based on a combination of multiplex PCR and biochip array hybridization. Illumina Human CNV370-Duo array was performed in a subsample of children population of SFS based on the protocol that has been described previously (Froguel, Ndiaye et al. 2012). The two polymorphisms of LSR were genotyped by LGC group as described before in the adults populations of SFS and ApoEurope population, while in the children population of SFS, these SNPs were not directly genotyped in the CNV370-Duo array, so, for the analysis, we have used proxies directly genotyped in our sample: rs2853332 (Rsquared: 0.965) for rs916147 and rs482324 (Rsquared: 1) for rs34259399.

The description of genotyping of the single-nucleotide polymorphism (SNP) of APOE rs429358 (Cys112Arg), and rs7412 (Arg158Cys) (commonly mentioned as APOE common polymorphism) were presented by Hixson and Vernier (Hixson and Vernier 1990). The alleles of APOE common variant are: Cys112/Cys158 (ε2), Cys112/Arg158 (ε3), and Arg112/Arg158 (ε4).

Table 3 presents the characteristics of the assessed genetic determinants of the 3 assessed molecules (VEGF, APOE, LSR)

Genome-wide DNA methylation profiling was performed by the Infinium HumanMethylation450 BeadChip (Illumina) (Bibikova, Barnes et al. 2011). Methylation ratio, referred to as beta value by Illumina’s software, is the proportion methylated / (methylated + unmethylated) CpGs.

Table 3: Characteristics of the assessed VEGF, LSR and APOE polymorphisms

Variant rsID Chr Allele 1 Allele 2 Near Gene Location rs114694170 5 T C MEF2C Intron rs34528081 6 T - VEGFA Intergenic

LOC10013235 rs6921438 6 A G Intergenic 4 rs1740073 6 T C C6orf223 Intergenic (proxy off rs4416670) rs6993770 8 A T ZFPM2 Intron rs7043199 9 A T VLDLR-AS1 Intron rs2375981 9 C G KCNV2 Intergenic (proxy of rs10738760) rs10761741 10 T G JMJD1C Intron rs4782371 16 T G ZFPM1 Intron rs2639990 18 T C ZADH2 Intron rs916147 19 A G LSR Intron rs429358 19 T C APOE exon rs7412 19 T C APOE exon

Statistical analysis

Hardy–Weinberg equilibrium was tested using a χ2-test.

The values of phenotypes of blood lipids such as LDL, HDL, TG, TC, APOE etc. were transformed if they were not normally distributed in the population. The associations of SNPs and the phenotypes were evaluated by linear regression models using the PLINK software (version 1.07), under the hypothesis of an additive model, where the minor allele is the reference allele. The models were adjusted for age, gender, BMI and other risk factors of chronic diseases.

The gene x gene interactions analyses were also performed by linear regression models under the hypothesis of an additive model. In case-control analysis, logistic regression models were used under additive model.

Furthermore, linear mixed-effects models were carried out for epigenetic analysis for DNA methylation. These models describe the relationship between a response variable and independent variables, adjusting for any relevant covariates (Pinherio and Bates. 2004). It consists of two parts, fixed effects, which include the conventional linear regression and random effects, which involve experimental units drawn at random from a population. Bonferroni correction methods (<0.05) were considered for defining the statistical significance (Benjamini and Hochberg 1995).

Publications - Results

Publication 1

Pharmacogenomic Challenges in Cardiovascular Diseases: Examples of Drugs and Considerations for Future Integration in Clinical Practice.

Chatelin J*, Stathopoulou MG*, Arguinano AA*, Xie T*, Visvikis-Siest S*. Curr Pharm Biotechnol. 2017;18(3):231-241. *equal contribution

In this review, we are presenting the challenges, advances and perspectives of the CVD medications and the application of pharmacogenomics in medical practice. There is also a focus on effectiveness and adverse effects of CVD drugs. As pharmacogenomics is the basis of personalisation of the treatment, its application in CVD medical practice requires the study of human genome in regards to drugs pharmacokinetics, pharmacodynamics, interactions and tolerance profile. The existing state–of–the–art of CVD drugs gives hopes for a future revolution in the drug development that will maximise cardiovascular patients benefit while decreasing their risks for adverse effects. The application of pharmacogenomics in the CVD medical practice is facing many methodological, technical, ethical, behavioral and financial issues, while cost-effectiveness is among the main prerequisite. The consideration of gene × gene × environment interactions and the inclusion of “omics” data in pharmacogenomic studies of CVD drugs will facilitate the generation of reliable results and will promote tailored treatments and new strategies of drug research and development.

100

101

102

103

104

105

Publication 2

A transnational collaborative network dedicated to the study and applications of the Vascular Endothelial Growth Factor–A in medical practice: The VEGF Consortium

Stathopoulou MG, Xie T, Ruggiero D, Chatelin J, Rancier M, Weryha G, Kurth MJ, Aldasoro Arguinano AA, Gorenjak V, Petrelis AM, Dagher G, Dedoussis G, Deloukas P, Lamont J, Marc J, Simmaco M, van Schaik RHN, Innocenti F, Merlin JL, Schneider J, Alizadeh BZ, Ciullo M, Seshadri S, Visvikis-Siest S, VEGF Consortium

Accepted for publication in Clinical Chemistry and Laboratory Medicine

VEGF is considered as a potential critical factor for regulating most important chronic diseases such as CVD, cancer, diabetes, chronic obstructive pulmonary disease (COPD), bone diseases and others. Thus, VEGF is a biomarker that may offer applicable information with very promising role in implementation of personalised medicine in several common chronic diseases. The Vascular endothelial growth factor European Genomic Federation – VEGF Consortium (www.vegfconsortium.org) was founded on June 2014 by an international group of researchers with an interest on VEGF and its implications in personalised medicine.

The Consortium is following a strategy of “systems biology” by integrating a variety of experimental, theoretical and computational methodologies for the development of clinically applicable VEGF “omics” tools in personalised medicine.

Among the results of the Consortium, two GWAS studies have associated 10 VEGF related polymorphisms that explain more than 50% of the individual variability of VEGF levels (Debette, Visvikis-Siest et al. 2011, Choi, Ruggiero et al. 2016).

106

Within the frames of the consortium, I have performed epistatic interactions analysis between the VEGF-related polymorphisms and the intermediate phenotypes for CVD in SFS, while 4 more populations participated in the analyses. A meta-analysis has been performed and a replication in UK Biobank is ongoing.

Population and analysis:

SFS unrelated adults from the second visit were included in this project (n=327). Furthermore, supposed healthy individuals from Cilento study (n=1374), from LifeLines cohort (n=13388), from ApoEurope (n=1706) and the hypertensive populations (n=945) were also analysed.

We have selected 15 SNPs combinations of the 10 VEGF-related polymorphisms to test against multiple CVD intermediate phenotypes: systolic blood pressure (SBP), diastolic blood pressure (DBP), pulse pressure (PP), BMI, obesity, central obesity, WC, waist/hip ratio, lipid accumulation product (LAP) index, visceral adiposity index (VAI), and 10-year risk for CVD based on SCORE charts (Conroy, Pyorala et al. 2003). The 15 combinations were selected based on correlations between them in the three populations: rs4416670× rs6921438 rs114694170× rs6921438 rs4782371× rs114694170 rs114694170 x rs7043199 rs114694170 x rs2375981 (rs10738760) rs114694170 x rs2639990 rs4782371x rs10761741

107

rs34528081 x rs6993770 rs2375981 (rs10738760) x rs6993770 rs2639990 x rs34528081 rs6921438 x rs2375981 (rs10738760) rs7043199 x rs4782371 rs10761741×rs6993770 rs6993770×rs4782371 rs263990×rs4782371

LAP was calculated based on the equations below:

 LAP for men = (WC [cm] - 65) × (TG concentration [mmol/L])  LAP for women = (WC [cm] - 58) × (TG concentration [mmol/L]) VAI was calculated based on the equations below:

 VAI for men = (WC/39.68+(1.88×BMI)) × (TG/1.03) × (1.31/HDL)  VAI for women = (WC/39.58+(1.89×BMI)) × (TG/0.81) × (1.52/HDL) The epistatic interactions were assessed using linear regression models with the PLINK software (version 1.07), under the hypothesis of an additive model. The meta-analysis was performed using the GWAMA software in a random effect correction method. The cut-off for significance was set at P<0.0033.

Results:

In SFS none of the 15 combinations reached the level of significance for any of the assessed phenotypes. Nevertheless, in the meta-analysis of all the populations, four significant interactions between 5 polymorphisms were identified for PP, SBP and VAI

108

(table 4). The interactions rs2375981*rs6993770 and rs2639990*rs114694170 were associated with increased levels of PP and SBP respectively. The interaction rs6993770*rs34528081 has a protective effect as it is associated with decreased SBP levels and the interaction rs114694170*rs2375981 was associated with decreased VAI values, thus lower visceral adiposity.

Conclusion:

In conclusion, in this preliminary study, we have identified significant epistatic interactions between VEGF-related SNPs and CVD intermediate phenotypes, thus we have validated a part of the working aims of VEGF consortium, through a genetic epidemiological approach of human populations. This study is ongoing as replication in UK biobank is currently being performed

Table 4. Significant epistatic interactions on CVD intermediate phenotypes in the VEGF Consortium meta-analysis.

Interaction beta P-value N_studies N_samples Effects PP rs2375981*rs6993770 0.001 3.75×10-4 5 17287 ++-++ SBP rs2639990*rs114694170 0.010 5.54×10-4 3 2391 ?-?++ SBP rs6993770*rs34528081 -0.003 1.04×10-3 3 2299 ?-?-- VAI rs114694170*rs2375981 -0.044 1.61×10-3 3 14931 ---

109

110

111

112

113

Publication 3

VEGF-related polymorphisms identified by GWAS and risk for major depression

Xie T*, Stathopoulou MG*, de Andrés F, Siest G, Murray H, Martin M, Cobaleda J, Delgado A, Lamont J, Peñas-LIedó E, LLerena A, Visvikis-Siest S.

Transl Psychiatry. 2017 Mar 7;7(3):e1055. *equal first authors

VEGF is recognised as a major inducer of angiogenesis and vasculogenesis and it has been shown to have effects in the CNS. Therefore, relationships between VEGF and mental diseases can be hypothesised. Therefore in this study we assessed the associations of 4 VEGF-related polymorphisms (rs10738760, rs6921438, rs6993770, rs4416670) with the risk of depression.

Population and analysis:

Depression patients (n = 437 subjects) and healthy controls from Spanish Caucasian (n = 477 subjects) origin were included in this study. Analysis was performed using the Generalized Multifactor Dimensionality Reduction GMDR0.9 software, which is a nonparametric and genetic model-free alternative to linear or logistic regression for detecting and characterizing nonlinear interactions among discrete genetic and environmental attributes.

114

Results:

Direct effects:

Among the four SNPs studied, only the rs4416670 was significantly associated with depression. This SNP was associated with an increased risk for depression (OR: 1.60; p = 0.01), while the minor allele C was more frequent in patients. No association with depression was found for the other three SNPs.

Epistatic interactions:

As only one SNP (rs4416670) was directly associated with depression, we have assessed only the epistatic interactions involving this SNP. The significance cut off value was set at P 0.05/3 (3 interaction tests) =0.016. No significant interaction was identified.

Conclusion:

In this project, we have detected common genetic regulation between genetic markers of VEGF and depression. The identification of SNPs that can affect the risk of developing depression and could hypothetically be implicated with treatment resistant depression, as well as the identification of the functional role of these SNPs, could allow the implementation of these findings in personalised risk prediction and hopefully targeted treatment of depression.

115

116

117

118

Publication 4

The role of epistatic interactions between 4 VEGF-related SNPs on osteoporosis

Chatelin J*, Xie T*, Kranjc T*, Stathopoulou MG, Toupance S, Rancier M, Vance D, Doherty L, Masson C, Murray H, Lamont J, Fitzgerald P, Weryha G, Benetos A, Marc J, Visvikis-Siest S. VEGF and osteoporosis; A VEGF Consortium collaborative study. * equal first authors

Under preparation

VEGF is a highly heritable molecule implicated in many chronic diseases such as bone diseases. Furthermore, murine and human osteoblastic cells produce VEGF acting in the paracrine regulation of blood vessels formation in bone tissue and in the autocrine regulation of osteogenesis. Depending on the results of VEGF Consortium, as far, that 10 single nucleotide polymorphisms (SNPs) are associated with VEGF plasma levels, we assumed that these SNPs could have a link with phenotypes of osteoporosis. We tested this hypothesis on Slovenian and ARTEOS populations of post-menopausal women by performing an epistatic interactions study.

Population and analysis:

The Slovenian and ARTEOS populations of post-menopausal women were used in this project. Associations between the polymorphisms and the bone mineral density (BMD) of lumbar spine (LsBMD) and femoral neck (FnBMD), the risk for osteoporosis and the age of menopause were tested. The package Model Based Multifactor Dimensionality Reduction in R was used for the analyses. Linear regression models were applied under additive model, adjusted for primary covariates

119

of age, BMI, and diagnosis for the associations with BMD. The meta-analysis was performed using the GWAMA software in a random effect correction method.

Results:

Sample characteristics are shown in table 5 and polymorphisms’ characteristics are presented in table 6.

Direct effects:

No significant association was identified.

Epistatic interactions:

The cut-off value for p-value was set to 0.002 for the associations with BMD and 0.008 for the associations with osteoporosis and age of menopause.

The epistasis between rs6921438 x rs1073876 is significantly associated with femoral neck bone mineral density (BMD-fn) and with T-score of femoral neck in the younger post-menopausal population. However, this finding is not replicated in the older post-menopausal women who are all osteoporotic. Also, no replication was found in ARTEOS population, which is also older in age, as shown in table 7(A). To be noted that in ARTEOS population, only total femur BMD (coltotalBMD) values were available. Furthermore this interaction could not be observed in the meta-analysis (table 7(B)) of the populations. The meta-analysis has revealed another significant interaction between rs6921438*rs6993770 and FnBMD, not identified in the level of individual studies.

For the phenotype age of menopause, several epistatic interactions were identified in different populations (table 8(A)). However, these interactions were not significant through the meta-analysis (table 8(B)). ARTEOS population was not included in this analysis because data on menopause age were not available. 120

No significant interactions were identified for osteoporosis.

Discussion-Conclusion:

We have detected common genetic regulation through epistasis between genetic markers of VEGF and BMD in post-menopausal populations. The identification of the interaction rs6921438 x rs1073876 could affect the femoral neck BMD in younger ages while in older women the effect is not identified. The meta-analysis has identified the rs6921438 x rs6993770 interaction that was not evident in the individual population analysis, probably due to increased power of meta-analysis.

In one previous study, we have identified that the interaction rs6921438 x rs6993770 was associated with levels of interleukin 6 (IL6) (Azimi-Nezhad, Stathopoulou et al. 2013). The link of osteoporosis and bone mass with IL6 is well established, especially in post-menopausal women (Zhao 2012). Therefore, as presented in figure 12, an explanation of the effect of the epistatic interaction between the rs6921438 x rs6993770 VEGF-related SNPs on FnBMD could be mediated by their effect on IL6 levels.

121

Table 5: Characteristics of post-menopausal women populations

Mean±SD

Younger Older ARTEOS post-menopausal post-menopausal (N=97) women (N=52) women (N=91) Age (years) 54.8±4.41 79.07±6.80 67.52±5.86 BMI (kg/m2) 27.11±4.72 27.16±4.76 25.50±3.87 BMD_fn (g/cm2) 0.76±0.12 0.58±0.14 BMD_ls (g/cm2) 0.96±0.14 0.85±0.23 -1.29±1.49 BMD col total NA NA -1.11±1.07 (g/cm2) Age of menopause 48.79±8.28 45.73±13.04 NA (years) Osteoporosis (%) 69.2 100 19.6

122

Table 6: Polymorphisms’ characteristics in the post-menopausal women populations

HWE MAF (minor allele frequency)

younger older ARTEOS younger older ARTEOS

post-menopausal post-menopausal post-menopausal post-menopausal

women women women women rs4416670 T/C 0.75 1 0.38 0.33 0.47 0.36 rs6993770 A/T 1 0.32 0.63 0.23 0.3 0.3 rs10738760 A/G 0.26 0.21 0.14 0.45 0.46 0.40 rs6921438 G/A 0.57 0.53 0.83 0.45 0.48 0.41

123

Table 7(A): Associations of epistatic interactions with FnBMD (P value cut off: 0.002)

Younger post-menopausal Older post-menopausal ARTEOS

BETA SE P BETA SE P BETA SE P

Fn BMD rs6921438*rs6993770 0.28 0.04 0.08 0.49 0.02 0.87 0.66 0.03 0.02 rs6921438*rs10738760 -0.58 0.02 0.0001 -0.15 0.03 0.44 0.02 0.03 0.08 rs6921438*rs4416670 -0.53 0.04 0.53 0.02 0.03 0.49 0.21 0.03 0.58 rs4416670*rs6993770 -0.1 0.03 0.72 -0.43 0.03 0.45 0.15 0.03 0.44 rs4416670*rs10738760 0.17 0.03 0.7 -0.13 0.02 0.54 0.18 0.03 0.82 rs10738760*rs6993770 0.06 0.03 0.65 0.57 0.03 0.51 0.24 0.03 0.08

T-score Fn rs6921438*rs6993770 0.298 0.357 0.143 0.483 0.209 0.111 0.5 0.247 0.006 rs6921438*rs10738760 -0.602 0.241 0.0005 -0.156 0.235 0.487 0.12 0.184 0.007

124

rs6921438*rs4416670 -0.599 0.353 0.078 0.039 0.229 0.262 0.206 0.235 0.982 rs4416670*rs6993770 0.006 0.278 0.733 -0.407 0.246 0.188 0.225 0.247 0.129 rs4416670*rs10738760 0.181 0.269 0.503 -0.117 0.194 0.077 0.015 0.211 0.322 rs10738760*rs6993770 0.046 0.328 0.371 0.571 0.233 0.476 0.04 0.21 0.04

Table 7(B): Meta-analysis of FnBMD results (P-value cut-off 0.002) fnBMD rs_number beta p-value n_studies n_samples effects rs6921438*rs6993770 0.478560 1.52*10-7 3 240 +++ rs6921438*rs10738760 -0.237169 0.223963 3 240 +-- rs6921438*rs4416670 -0.099173 0.621541 3 240 +-+ rs4416670*rs6993770 -0.126667 0.450774 3 240 +-- rs4416670*rs10738760 0.072440 0.520433 3 240 ++- rs10738760*rs6993770 0.290000 0.052154 3 240 +++

125

T-scorefn rs_number beta p-value n_studies n_samples effects rs6921438*rs6993770 0.458114 0.001668 3 240 +++ rs6921438*rs10738760 -0.191747 0.367834 3 240 --+ rs6921438*rs4416670 -0.047763 0.817992 3 240 -++ rs4416670*rs6993770 -0.062221 0.746394 3 240 +-+ rs4416670*rs10738760 -0.004303 0.972773 3 240 +-+ rs10738760*rs6993770 0.233769 0.206243 3 240 +++

Table 8(A): Associations of interactions with age of menopause (P value cut off: 0.008)

Younger post-menopausal Older post-menopausal

BETA SE P BETA SE P rs6921438*rs6993770 -0.26 1.66 0.49 -0.31 1.14 2.9*10-7

126

rs6921438*rs10738760 -0.25 1.26 0.0302 0.11 1.26 0.96 rs6921438*rs4416670 -0.23 1.86 4.7*10-14 0.04 1.26 0.029 rs4416670*rs6993770 -0.16 1.34 0.003 0.63 1.32 0.01 rs4416670*rs10738760 -0.04 1.36 0.01173 0.43 1.03 0.98 rs10738760*rs6993770 -0.32 1.65 0.44 0.02 1.27 2.2*10-7

Table 8(B): Meta-analysis of age of menopause results (P-value cut-off 0.08) rs_number beta p-value n_studies n_samples effects rs6921438*rs6993770 -0.293976 0.754417 2 143 -- rs6921438*rs10738760 -0.070000 0.937357 2 143 -+ rs6921438*rs4416670 -0.044929 0.965632 2 143 -+ rs4416670*rs6993770 0.240940 0.797780 2 143 -+ rs4416670*rs10738760 0.258681 0.752734 2 143 -+ rs10738760*rs6993770 -0.106490 0.915710 2 143 -+

127

Figure 12. The effect of rs6921438 x rs6993770 on FNBMD could be mediated by IL6.

128

Publication 5

The effect of LSR functional polymorphisms on blood lipid traits and the role of the epistatic interactions between 2 candidate genes: APOE and LSR

Xie T*, Stathopoulou MG*, Akbar S, Oster T, Siest G, Yen FT, Visvikis-Siest S.

Effect of LSR polymorphism on blood lipid levels and age-specific epistatic interaction with the APOE common polymorphism *equal first authors

Submitted to Clinical Genetics

LSR is an apolipoprotein (Apo) B and E receptor that participates in the removal of

TG-rich lipoproteins. APOE is an important risk factor for CVD, especially in type III

HLP. Furthermore, LSR gene is located upstream of the APOE, thus we suppose that LSR is a novel candidate gene for CVD and that possible epistatic interactions with APOE polymorphism could also be hypothesised.

Population and analysis:

We used SFS unrelated adults (n=432) and unrelated children (n=328) and ApoEurope cohort for this project.

The epistatic interactions analysis was carried out by EPISNP1 (a random population with unrelated individuals) mode of epiSNP (version 4.2) for adults, children and replication population. Three different combinations of alleles of APOE were created and coded as: 2= ε2/2, 1= ε3/2 and ε2/4, and 0= ε3/3, ε3/4, ε4/4 for ε2 allele and 2= ε4/4, 1=

ε2/4, ε3/4 and 0= ε2/2, ε3/2, ε3/3for ε4 allele). Level of significance has been set to 0.008 for direct associations as only 1 polymorphism has been tested in 6 phenotypes and to 129

0.004 for epistatic interactions (2 interactions between one SNP of LSR with two groups of APOE alleles in 6 phenotypes) (Bonferroni correction)

Results:

Sample characteristics are shown in table 9 and polymorphisms’ characteristics are presented in table 10.

Direct effects:

A direct nominally significant effect of LSR rs916147 was found in both populations

(adults and children) for plasma apolipoprotein B levels, while a negative association was also observed between this SNP and LDL in children only (Table 11). No other association of rs916147 was significant for other lipid traits and no significant effect was detected for the other lipid traits (data not shown) in the replication population; the

ApoB levels were not available for this population.

Epistatic interactions:

Epistatic interactions were identified between the LSR rs916147 and APOE ε2 allele in adults for TC, APOE and LDL levels (Table 12(A)), while for the ε4 group, no significant interaction was observed (data not shown). In children, no significant interaction was observed for LSR SNP rs916147 and APOE.

In order to underline the importance of the epistasis between the assessed epistatic interactions of LSR SNP with APOE, we have calculated the direct effect of APOE variant on blood lipids (table 12(B)). These data revealed that APOE ε2 alone in adults was related with decreased levels of TC, APOE and LDL, while its epistatic interaction with LSR rs916147 resulted in increased levels of the above lipid traits. This led us to

130

conclude that the epistatic interaction of the LSR SNP rs916147 with APOE ε2 reversed the direct effect of APOE ε2 on TC, APOE and LDL levels.

In the replication population, similar effects were observed for TC and APOE levels

(Table 12(A) & (B)).

Conclusion:

In conclusion, in this project, through a genetic epidemiological approach of human populations, we have validated the working hypothesis based on previously obtained mechanistic results from both cell and animal models, which lead us to propose a potential role of LSR in the regulation of lipid metabolism in humans. We would therefore propose that LSR could be considered as a novel therapeutic target for treatment of hyperlipidemia, a significant risk factor of CVD. This study represents the first report demonstrating that LSR SNP rs916147 can affect lipid levels in humans and reveals the existence of complex molecular links between LSR and APOE that merit further investigation which could provide new insights in the mechanisms underlying type III HLP.

131

Table 9: Characteristics of the SFS adults and children and the ApoEurope replication population

Children SFS Adults SFS Adults replication

Total (N=328) Total (N=432) Total (N=1744)

Mean SD Mean SD Mean SD

Age (years) 15.39 1.64 44.85 4.36 42.00 9.87

Body Mass Index (kg/m2) 19.71 2.26 24.98 3.61 26.22 4.68

Cholesterol (mmol/L) 4.42 0.80 5.72 0.93 5.72 1.22

Triglycerides (mmol/L) 0.88 0.46 1.24 0.70 1.48 1.15

High density lipoprotein 1.52 0.35 1.60 0.44 1.53 0.57 (mmol/L)

Low density lipoprotein 2.72 0.72 3.51 0.99 3.51 1.14 (mmol/L)

Apolipoprotein B (g/L) 0.72 0.17 1.01 0.22 NA NA

Apolipoprotein E (mg/L) 38.56 10.47 40.87 10.99 44.90 14.04

132

Table 10: Polymorphisms’ characteristics of the SFS adults and children and the ApoEurope replication population

Children SFS Adults SFS Adults replication

e Gen. Gen. Gen. MAF Chr SNP MAa MAFb HWEc MAFb HWEc MAFb HWEc Rated Rated Rated 1KG

19 rs916147 G 0.41 0.15 98.8 0.40 0.19 99.3 0.39 0.18 92.3 0.37

19 rs429358 C 0.09 1 98.5 0.12 0.82 100 0.12 1 93.7 0.09

19 rs7412 T 0.09 0.51 98.5 0.10 0.80 100 0.06 0.82 93.7 0.06 a Minor allele, b Minor allele Frequency, c Hardy-Weinberg equilibrium P-value, d Genotyping rate, e Minor allele Frequency from the European population (CEU) in 1000 Genomes project

133

Table 11: Direct effects of rs916147 LSR gene polymorphism

rs916147

Children SFS Adults SFS Adults replication

beta±SEa P beta±SEa P beta±SEa P

Cholesterol (mmol/L) -0.111±0.058 0.059 -0.003±0.005 0.484 0.04±0.003 0.98

Triglycerides (mmol/L) 0.027±0.031 0.396 -0.003±0.014 0.836 -0.006±0.07 0.127

High density lipoprotein (mmol/L) -0.003±0.025 0.896 0.002±0.007 0.755 -0.013±0.005 0.946

Low density lipoprotein (mmol/L) -0.112±0.053 0.035 -0.004±0.007 0.543 0.025±0.005 0.431

Apolipoprotein B (g/L) -0.029±0.012 0.020 -0.013±0.006 0.042 NA NA

Apolipoprotein E (mg/L) -0.157±0.740 0.832 0.005±0.008 0.498 0.027±0.007 0.697 a Standard Error

134

Table 12 (A): Significant epistatic interactions of rs916147 SNP of LSR with APOE polymorphism

P P P beta±SEa beta±SEa beta±SEa interaction interaction interaction Interaction Traits log(10) children adults adults children adults adults (SFS) (SFS) (replication) (SFS) (SFS) (replication)

Triglycerides -0.211±0.052 0.999 0.243±0.032 0.051 0.003±0.022 0.471

rs916147 Cholesterol 0.001±0.02 0.999 0.114±0.011 0.777×10-8 0.017±0.009 0.272×10-5 (rs2853332) -3 -5 × Apolipoprotein E -0.128±0.029 0.999 0.125±0.019 0.639×10 0.005±0.006 0.408×10

APOE ε2 Low density -0.083±0.029 0.999 0.059±0.017 0.531×10-3 0.020±0.016 0.479 lipoprotein

135

High density 0.165±0.027 0.999 0.019±0.017 0.607 -0.002±0.017 0.842 lipoprotein

Apolipoprotein B -0.113±0.026 0.999 -0.036±0.014 0.513 NA NA a Standard Error

Table 12(B): Direct effect of APOE genotypes on the above lipids traits

beta±SE a P beta±SE a P beta±SE a P

Locus Traits log(10) children children adults adults adults adults

(SFS) (SFS) (SFS) (SFS) (replication) (replication)

Triglycerides 0.087±0.025 0.257 0.057±0.023 0.093 -0.009±0.016 0.625 APOE ε2 Cholesterol -0.257±0.009 0.439×10-4 -0.101±0.008 0.003 -0.124±0.006 0.614×10-5

136

Apolipoprotein E 0.339±0.014 0.627×10-2 0.374±0.013 0.766×10-26 0.123±0.009 0.339×10-8

Low density -0.317±0.014 0.257×10-7 -0.206±0.012 0.111×10-4 -0.141±0.012 0.489×10-5 lipoprotein

High density 0.030±0.013 0.799 0.099±0.012 0.313×10-3 0.018±0.012 0.002 lipoprotein

Apolipoprotein B -0.347±0.012 0.259×10-8 -0.230±0.001 0.147×10-5 NA NA a Standard Error

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

Publication 6

The role of the epistatic interactions between APOE and LSR genetic variants on Alzheimer’s disease.

Xie T*, Akbar S*, Stathopoulou MG*, Oster T, Masson C, Yen FT, Visvikis-Siest S. The association of the epistatic interaction of APOE and LSR genetic variants in Alzheimer’s disease. * equal first authors

Under submission

The APOE common polymorphism (ε4 allele) is the strongest risk factor for the late-onset AD. Human APOE gene is located on chromosome 19q13.1, a region linked to AD. Another candidate gene mapped in the same region is LSR, which encodes the lipolysis stimulated lipoprotein receptor. As an APOE-receptor, LSR plays an important role in the removal of triglyceride-rich lipoproteins and therefore regulates lipid homeostasis in both periphery and central nervous system. Therefore, we assumed that LSR might be also implicated in AD directly or by interaction with APOE common polymorphism.

Population and analysis:

In this project, the Alzheimer’s diseases case-control European population was used (142 AD patients and 63 controls).

The analysis was performed using the package Model Based Multifactor Dimensionality Reduction (mbmdr) in R. Three different combinations of allele of

APOE ε4 (2= ε4/4, 1= ε2/4, ε3/4 and 0= ε2/2, ε3/2, ε3/3) were defined. Cut-off level for

158

significance was set to 0.025 for direct effects associations and to 0.008 for epistatic interactions.

Results:

Sample characteristics are shown in table 13. The polymorphisms’ characteristics are shown in table 14.

Direct effects:

No direct effects were identified between LSR SNPs and risk for AD.

Epistatic interactions:

Our data indicate significant epistatic interactions between APOE and the two LSR SNPs, rs34259399 and rs916147 (Table 15). Furthermore, our analysis revealed that the interactions between the LSR SNPs and the APOE common polymorphisms of non-e4 allele combination (ε2/2, ε3/2, ε3/3) lead to an increased risk of AD (Table 16 & 17).

Conclusion:

This is the first report that demonstrates the potential involvement of the LSR genetic variants through their epistatic interaction with APOE in AD. Our results indicate the presence of significant interactions between the novel candidate gene, LSR, and the APOE common polymorphism, the greatest well-known risk factor for AD. The analysis revealed that the non-e4 allele combination (ε2/2, ε3/2, ε3/3) interacted significantly with both LSR SNPs, indicating that the non-e4 allele is associated with an increased disease risk. Due to the role of LSR as an APOE-receptor in lipid homeostasis, the chromosomal location of LSR gene in chromosome 19, as well as the

159

significant APOE-LSR interactions in AD population, LSR represents an interesting candidate to predict an individual’s susceptibility to AD and needs to be explored further in order to understand the detailed mechanisms of molecular interactions in AD pathology.

Table 13: AD patients and controls characteristics

Patients (142) Control (63) Mean sd Mean sd Age (years) 73.16 8.54 70.41 2.56

Cholesterol (mmol/L) 5.49 1.32 5.61 1.31 Triglycerides (mmol/L) 1.42 0.80 1.54 0.66 Apolipoprotein E (mg/L) 43.44 17.32 48.23 12.51

Table 14: Polymorphisms’ characteristics of AD patients and controls

Patients Control

Gen. Gen. Chr SNP MAa MAFb HWEc MAFb HWEc Rated Rated

19 rs916147 G 0.35 5.6 86.7 0.37 0.0014 95.2

19 rs34259399 A 0.14 0.17 88.7 0.1 0.71 98.4

19 rs429358 C 0.34 0.03 100 0.1 0.62 100

19 rs7412 T 0.05 0.44 100 0.06 0.2 100 a Minor allele, b Minor allele Frequency, c Hardy-Weinberg equilibrium Chi-squared, d Genotyping rate.

160

Table 15: Significant epistatic interactions of rs34259399 and rs916147 SNPs of LSR with APOE

Interaction beta SE p-value rs34259399 x -0.95 1.19 0.00002 APOE non-e4 allele

rs916147 x -0.83 0.56 0.0068 APOE non-e4 allele

Table 16: Significant epistatic interactions between APOE and rs34259399

SNP1 SNP2 Case Control beta p-value Cat*

APOE non-e4 allele rs34259399

ε2/2, ε3/2, ε3/3 C/C 45 41 -1.44 0.00002 L**

*Category: Predicted risk category for the genotype.

**L : Low risk

Table 17: Significant epistatic interactions between APOE and rs916147

SNP1 SNP2 Case Control beta p-value Cat*

APOE non-e4 allele rs916147

ε2/2, ε3/2, ε3/3 G/A 25 23 -0.98 0.0068 L

*Category: Predicted risk category for the genotype.

**L : Low risk

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

Publication 7

Epigenome-wide association study on central obesity phenotypes.

Xie T*, Gorenjak V*, Stathopoulou M*, Dade S, Vance D, Doherty L, Marouli E, Masson C, Murray H, Lamont J, Fitzgerald P, Deloukas P, Visvikis-Siest S. Epigenome-wide association study in healthy individuals shows significant correlation with cg16170243 and waist circumference level. *equal first authors

Under submission

Epigenetic changes are molecular pathways that are modulating the genetic expression thus are able to affect a particular phenotype and are affecting physiological mechanisms as well as pathophysiology of many diseases. Therefore, we performed an epigenome-wide association study (EWAS) on central obesity phenotypes of healthy individuals aiming to study the effect of the observed DNA methylation changes and to investigate the possible involved mechanisms.

Populations and analysis:

Healthy adults and children from the first recruitment of SFS (220 subjects from 75 families) were used in this project. The package minfi and CpGassoc (a linear mixed-effects model) in R were carried out for epigenetic analysis.

177

Results:

Sample characteristics are shown in table 18.

Our study showed a significant association of cg16170243 with WC (table 19). Cg16170243 corresponds to a 50 bp-length human methylation oligoprobe located on chromosome 18 between 46 759 502 bp and 46 759 551 bp. This methylation site is located in AC090241.2 gene. This gene has 3 splice variants and cg16170243 is located both in intronic zone of AC090241.2-202 variant and in 5’ of AC090241.2-203 variant.

AC090241.2-202 variant is an “antisense” biotype transcript meaning that it overlaps the genomic span of a protein-coding locus on the opposite strand. The overlapping gene on the opposite strand is ST8SIA5 (Alpha-2,8-sialyltransferase 8E) coding for a sialyltransferase enzyme involved in the synthesis of gangliosides GD1c, GT1a, GQ1b and GT3 from GD1a, GT1b, GM1b and GD3 respectively (Kim, Kim et al. 1997). Thus cg16170243 methylation of AC090241.2 antisense gene could modify ST8SIA5 expression and consequently the synthesis of gangliosides (Figure 13).

Conclusion:

As previously shown, an inadequate ganglioside expression in mediobasal hypothalamic neurons deregulates insulin signaling (Herzer, Meldner et al. 2015), which may lead to obesity. Thus, a potential insulin signaling pathway could be hypothesised to be implicated in the explanation of the association between the specific probe methylation and WC, through modifying antisense transcription of AC090241.2 to affect ST8SIA5 transcription and its enzymatic activity, thereby subsequent gangliosides. These mechanisms need to be verified but it is known that the deregulated insulin signaling may significantly contribute to systemic insulin resistance in central obesity (Hardy, Czech et al. 2012)

178

Table 18: Characteristics of the SFS population

Total Adults (116) Children (95) Men (105) Women (106)

mean sd mean sd mean sd mean sd mean sd

Age (years) 28.17 14.83 40.48 7.53 13.15 2.58 27.09 15.14 29.24 14.37

Body Mass Index (kg/m2) 21.52 4.0 24.06 3.22 18.43 2.31 21.77 4.17 21.27 3.79

Waist circumference (cm) 72.92 11.82 80.11 10.28 64.15 6.29 76.62 13.08 69.26 8.96

Hip (cm) 90.38 11.12 97.13 6.27 82.14 10.09 89.38 11.76 91.37 10.3

Waist-to-hip ratio 0.81 0.08 0.82 0.08 0.79 0.06 0.86 0.07 0.76 0.05

Table 19: Association of methylation site with waist circumference level

Phenotype CpG Name Chr Effect size SE P-Value FDR Bonferroni

waist AC090241. cg16170243 18 2.32 0.41 0.13 × 10-8 0.048 0.048 circumference 2

179

Figure 13: cg16170243 could modify the aberrant transcription of AC090241.2 locus and repress its expression, which can lead to an antisense transcript that overlaps with ST8SIA5 gene.

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

Publication 8

Epigenome-wide association study on blood lipids.

Xie T*, Gorenjak V*, Stathopoulou M*, Marouli E, Vance D, Doherty L, Masson C, Murray H, Lamont J, Fitzgerald P, Deloukas P, Visvikis-Siest S. EWAS of blood lipids in healthy population from STANISLAS Family Study (SFS). *equal first authors

Under submission

Epigenetic mechanisms might be implicated in the regulation of interindividual variability of lipid levels and thus may play an important role to the cardiovascular risk profile. We have performed an EWAS on blood lipid levels: high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglycerides, and total cholesterol.

Populations and analysis:

201

Results:

Sample characteristics are shown in table 20.

Four probes showed statistically significant associations with triglyceride levels (FDR p value < 0.05), including individual probes in TMEM150A (Transmembrane Protein 150A) on chromosome 2, PRKAG2 (protein kinase AMP-activated non-catalytic subunit gamma 2) on chromosome 7, KREMEN2 (Kringle Containing Transmembrane Protein 2) on chromosome 16 and WDR83OS (WD Repeat Domain 83 Opposite Strand) on chromosome 19 (table 21). Among them, PRKAG2 and KREMEN2 can cause the setup of diverse diseases. The probe cg08897188 is located in 5’ UTR of PRKAG2 gene in regulating area for gene transcription, while cg04580029 is located in transcription start site (TSS) of KREMEN2 gene (1 exon). The other two probes were detected near the TMEM150A and WDR83OS genes.

Conclusion:

The mutations in PRKAG2 cause familial hypertrophic cardiomyopathy (HCM) (Blair, Redwood et al. 2001). However, Arad et al. revealed that the PRKAG2 mutations affect the novel myocardial metabolic storage disease rather than HCM, even though it can be presented as a primary cardiomyopathy. Furthermore, the mutations in LAMP2 (lysosomal-associated membrane protein) or PRKAG2 lead to glycogen-storage cardiomyopathy that resembles the phenotype of HCM (Arad, Benson et al. 2002, Arad, Maron et al. 2005). Arrhythmogenic right ventricular cardiomyopathy / dysplasia (ARVC/D) and HCM are the two most common conditions leading to sudden cardiac death in athletes, nevertheless, those two diseases are caused by mutations in different genes. Interestingly, our results showed that since the DNA methylation site cg08897188 has a specific potential of a transcription factor that is able to bind MZF1 (Myeloid zinc finger 1 factors, associated with ARVC/D), with a similarity score of 0.992 (https://www.genomatix.de), thereby could link the interrelationship between the 202

two diseases. Therefore, our findings suggest that PRKAG2 might be a key candidate that can link ARVC/D and HCM, since it is able to bind MZF1.

The low-density lipoprotein-related receptor 6 (LRP6) is a co-receptor for WNT that transmits the canonical Wnt/beta-catenin signaling cascade, which has been associated with Alzheimer’s disease (De Ferrari, Papassotiropoulos et al. 2007) and coronary artery disease (CAD) (Mani, Radhakrishnan et al. 2007, Sarzani, Salvi et al. 2011). Furthermore, Dickkopf WNT Signaling Pathway Inhibitor 1 (DKK1) is an antagonistic inhibitor of the WNT signaling pathway that acts by isolating the LRP6 co-receptor. KREMEN proteins cooperated with DKK1 to regulate Wnt/beta-catenin signaling (Mao, Wu et al. 2002), including KREMEN1 and KREMEN2. Interestingly, our results show that KREMEN2-related CpG site cg04580029 was significantly associated with TG in a healthy population, which is located in a TSS of KREMEN2. Therefore, our results suggest that KREMEN2 could be affecting the risk for AD and CAD by blocking Wnt/beta-catenin signaling thereby disassociating/inhibiting of the LRP6–DKK1 complex, thus it could be proposed as a novel therapeutic target. Nevertheless, the extended underlying mechanisms still need to be elucidated.

The associations of the probes near the TMEM150A and WDR83OS genes are novel and no associations with mechanisms and diseases are known for these genes.

At present, our findings represent the first report demonstrating that the association between DNA methylation and lipid TG could affect the risk for different diseases in humans and reveal the existence of complex links between DNA methylation and diseases that merit further investigation.

203

Table 20: Characteristics of the SFS population

Total Adults (116) Children (95) Men (105) Women (106)

mean sd Mean sd mean sd mean sd mean sd

Age (years) 28.17 14.83 40.48 7.53 13.15 2.58 27.09 15.14 29.24 14.37

Body Mass Index (kg/m2) 21.52 4.0 24.06 3.22 18.43 2.31 21.77 4.17 21.27 3.79

Cholesterol (mmol/L) 5.24 1.0 5.7 0.96 4.68 0.74 5.12 1.04 5.37 0.95

Triglycerides (mmol/L) 0.86 0.49 0.98 0.56 0.72 0.32 0.91 0.55 0.82 0.41

High density lipoprotein 1.44 0.41 1.45 0.42 1.43 0.39 1.34 0.38 1.54 0.42 (mmol/L)

Low density lipoprotein 3.63 0.97 4.06 0.96 3.11 0.68 3.59 1.02 3.67 0.91 (mmol/L)

204

Table 21: Association of methylation values with TG level.

Gene Name Chr Effect size SE P-Value FDR cg21101051 TMEM150A 2 -0.34 0.06 0.41 × 10-6 0.049 cg08897188 PRKAG2 7 -0.49 0.09 0.52 × 10-6 0.049 cg04580029 KREMEN2 16 0.44 0.08 0.53 × 10-6 0.049 cg25263559 WDR83OS 19 -0.31 0.06 0.42 × 10-6 0.049

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

General Discussion

222

In this thesis, we aimed to take advantage of powerful approaches, including candidate genes, the study of epistatic interactions, gene-environment interactions and DNA methylation to decipher the biological determinants of common molecules and pathways that are implicated in different chronic diseases. With those methods, we have demonstrated:

1. The importance of epistatic interactions between VEGF-related polymorphisms affecting cardiovascular intermediate phenotypes in an international consortium, the VEGF Consortium

2. The polymorphism rs4416670, a GWAS-identified polymorphism linked to VEGF levels and CVD risk factors, to be associated with increased risk for depression.

3. Common genetic regulation between genetic markers of VEGF and osteoporosis in post-menopausal populations.

4. One functional LSR polymorphism interaction with the APOE common polymorphism reversing the ε2 allele protective effect on lipid levels in an age-specific manner.

5. Interactions between LSR and APOE representing an interesting mechanism that may predict an individual’s susceptibility to AD, which needs to be further explored in order to understand the detailed mechanisms of molecular interactions in AD pathology.

6. An association between DNA methylation probes and WC through modifying antisense transcription of AC090241.2 that may affect ST8SIA5 transcription and its enzymatic activity, thereby gangliosides, thus suggesting a potential insulin signaling pathway involved in the related mechanism.

7. An association between DNA methylation probes and lipid TG able to affect different diseases in humans revealing the existence of complex links between 223

DNA methylation and diseases that merit further investigation. This could provide new insight in the mechanisms underlying CVD and contribute to new therapeutic possibilities.

In particular, we have demonstrated significant inter-relationships between candidate genes through epistatic interactions. Epistasis can be a powerful tool in the study of human diseases. In the present study, our results propose potential inter-relationships that may affect the risk of type III HLP and AD, through the genetic interactions between LSR and APOE gene polymorphisms. Thus, we could propose that LSR might act as a link between type III HLP and AD. The next step is to explore and understand the underlying mechanisms of the molecular interactions explaining these associations in type III HLP and AD pathology.

In addition, the associations between DNA methylation and WC and TG levels suggests that epigenetic modifications could provide new insights in the mechanisms underlying chronic diseases and contribute to new therapeutic possibilities. Those epigenetic modifications could alter the molecular pathways that are modulating the expression of a genotype into a particular phenotype and by that affecting physiological mechanisms as well as pathophysiology of many diseases such as CVD, depression and osteoporosis, which are also affected by VEGF related SNPs.

According to the wide diversity of functions of VEGF, results previously reported by the VEGF Consortium partners and our findings, VEGF could be proposed as a potential biomarker and, through specific interaction profiles, regulator of the development of CVD, depression, and bone diseases.

Taken together, our findings show the interrelationship between chronic diseases and genetic variants. However, there are still some limitations such as further replications of results in healthy and pathological populations, and the underlying mechanisms between the genetic variants and diseases that need to be elucidated.

224

General conclusion

During the three years of the thesis, I have investigated through different methodologies the effect of epistatic interactions between polymorphisms identified by GWAS and ‘candidate gene’ approaches on several chronic diseases. Our strategy revealed significant interactions with all assessed phenotypes that may be differentiated by age.

More precisely, our findings highlight complex inter-relationships between biomarkers like VEGF, LSR and APOE that through direct effects, epistatic interactions, epigenetic modifications and age could play a central role in diseases stratification based on individuals’ molecular “make-up”. The figure 14 presents a summary of the results identified through the projects and publications of the thesis and demonstrates the links between the findings and the proposed mechanisms that we propose a conclusion of this thesis (Figure 14).

In the following steps, all the results of this thesis will be explored using functional analysis in order to understand the detailed mechanisms and molecular interactions underlying the findings.

225

Figure 14. Summary of the identified links between the selected chronic diseases and the targeted markers through investigation of epistatic interactions and mDNA – links between all the projects and publications of the thesis.

226

References abcam. "Histone modifications: a guide." from http://www.abcam.com/epigenetics/histone-modifications-a-guide.

Ambrose, J. A. and R. S. Barua (2004). "The pathophysiology of cigarette smoking and cardiovascular disease: An update." Journal of the American College of Cardiology 43(10): 1731-1737.

American Psychiatric Association (2013). "Diagnostic and Statistical Manual of Mental Disorders." 5th Edition: DSM‑5 (American Psychiatric Association, 2013).

APA (1994). "Diagnostic and Statistical Manual of Mental Disorders." American Psychiatric Association: Washington, DC.

Arad, M., D. W. Benson, A. R. Perez-Atayde, W. J. McKenna, E. A. Sparks, R. J. Kanter, K. McGarry, J. G. Seidman and C. E. Seidman (2002). "Constitutively active AMP kinase mutations cause glycogen storage disease mimicking hypertrophic cardiomyopathy." J Clin Invest 109(3): 357-362.

Arad, M., B. J. Maron, J. M. Gorham, W. H. Johnson, Jr., J. P. Saul, A. R. Perez-Atayde, P. Spirito, G. B. Wright, R. J. Kanter, C. E. Seidman and J. G. Seidman (2005). "Glycogen storage diseases presenting as hypertrophic cardiomyopathy." N Engl J Med 352(4): 362-372.

Austin, M. A., J. L. Breslow, C. H. Hennekens, J. E. Buring, W. C. Willett and R. M. Krauss (1988). "Low-density lipoprotein subclass patterns and risk of myocardial infarction." Jama 260(13): 1917-1921.

Australian Institute of Health and Welfare. (2016). from http://www.aihw.gov.au/chronic-diseases/comorbidity/.

Azimi-Nezhad, M., M. G. Stathopoulou, A. Bonnefond, M. Rancier, A. Saleh, J. Lamont, P. Fitzgerald, N. C. Ndiaye and S. Visvikis-Siest (2013). "Associations of vascular endothelial growth factor (VEGF) with adhesion and inflammation molecules in a healthy population." Cytokine 61(2): 602-607.

Bachorik, P. S. and J. W. Ross (1995). "National Cholesterol Education Program recommendations for measurement of low-density lipoprotein cholesterol: executive summary. The National Cholesterol Education Program Working Group on Lipoprotein Measurement." Clin Chem 41(10): 1414-1420. 227

Bastard, J.-P., C. Jardel, J. Delattre, B. Hainque, E. Bruckert and F. Oberlin (1999). "Evidence for a Link Between Adipose Tissue Interleukin-6 Content and Serum C-Reactive Protein Concentrations in Obese Subjects." Circulation 99(16): 2219c-2222.

Bates, D. O., T. G. Cui, J. M. Doughty, M. Winkler, M. Sugiono, J. D. Shields, D. Peat, D. Gillatt and S. J. Harper (2002). "VEGF165b, an inhibitory splice variant of vascular endothelial growth factor, is down-regulated in renal cell carcinoma." Cancer Res 62(14): 4123-4131.

Benjamini, Y. and Y. Hochberg (1995). "Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing." Journal of the Royal Statistical Society. Series B (Methodological) 57(1): 289-300.

Bibikova, M., B. Barnes, C. Tsan, V. Ho, B. Klotzle, J. M. Le, D. Delano, L. Zhang, G. P. Schroth, K. L. Gunderson, J. B. Fan and R. Shen (2011). "High density DNA methylation array with single CpG site resolution." Genomics 98(4): 288-295.

Biomarkers Definitions Working, G. (2001). "Biomarkers and surrogate endpoints: Preferred definitions and conceptual framework." Clinical Pharmacology & Therapeutics 69(3): 89-95.

Bird, A. (2002). "DNA methylation patterns and epigenetic memory." Genes Dev 16(1): 6-21.

Blair, E., C. Redwood, H. Ashrafian, M. Oliveira, J. Broxholme, B. Kerr, A. Salmon, I. Ostman-Smith and H. Watkins (2001). "Mutations in the gamma(2) subunit of AMP-activated protein kinase cause familial hypertrophic cardiomyopathy: evidence for the central role of energy compromise in disease pathogenesis." Hum Mol Genet 10(11): 1215-1220.

Blom, E. S., P. Holmans, S. Arepalli, O. Adighibe, M. L. Hamshere, M. Gatz, N. L. Pedersen, A. L. Bergem, M. J. Owen, P. Hollingworth, A. Goate, J. Williams, L. Lannfelt, J. Hardy, F. Wavrant-De Vrieze and A. Glaser (2008). "Does APOE explain the linkage of Alzheimer's disease to chromosome 19q13?" Am J Med Genet B Neuropsychiatr Genet 147b(6): 778-783.

Boerwinkle, E., S. Visvikis, D. Welsh, J. Steinmetz, S. M. Hanash and C. F. Sing (1987). "The use of measured genotype information in the analysis of quantitative phenotypes in man. II. The role of the apolipoprotein E polymorphism in determining levels, variability, and covariability of cholesterol, betalipoprotein, and triglycerides in a sample of unrelated individuals." Am J Med Genet 27(3): 567-582. 228

Brandi ML, C.-O. P. (2006). "Vascular biology of the skeleton." Journal of Bone and Mineral Research 21: 183-192.

Braun, C., T. Bschor, J. Franklin and C. Baethge (2016). "Suicides and Suicide Attempts during Long-Term Treatment with Antidepressants: A Meta-Analysis of 29 Placebo-Controlled Studies Including 6,934 Patients with Major Depressive Disorder." Psychother Psychosom 85(3): 171-179.

Brulliard, M., V. Ogier, L. Mejean, L. Bonnard, B. Bihain and F. Yen (2008). "A polymorphism of the lipolysis-stimulated lipoprotein receptor gene is associated with elevated plasma triglycerides and LDL cholesterol in obese subjects. ." ATVB 28(6): E136.

Brunner, D., S. Altman, K. Loebl, S. Schwartz and S. Levin (1977). "Serum cholesterol and triglycerides in patients suffering from ischemic heart disease and in healthy subjects." Atherosclerosis 28(2): 197-204.

Cao, L., X. Jiao, D. S. Zuzga, Y. Liu, D. M. Fong, D. Young and M. J. During (2004). "VEGF links hippocampal activity with neurogenesis, learning and memory." Nat Genet 36(8): 827-835.

Cedar, H. and Y. Bergman (2009). "Linking DNA methylation and histone modification: patterns and paradigms." Nat Rev Genet 10(5): 295-304.

Centers for Disease, C., Prevention, P. National Center for Chronic Disease, P. Health, S. Office on and Health (2010). Publications and Reports of the Surgeon General. How Tobacco Smoke Causes Disease: The Biology and Behavioral Basis for Smoking-Attributable Disease: A Report of the Surgeon General. Atlanta (GA), Centers for Disease Control and Prevention (US).

Chang, T. Y., Y. Yamauchi, M. T. Hasan and C. C. Chang (2017). "Cellular Cholesterol Homeostasis in Alzheimer's Disease." J Lipid Res.

Chatelin, J., M. G. Stathopoulou, A. A. Arguinano, T. Xie and S. Visvikis-Siest (2017). "Pharmacogenomic Challenges in Cardiovascular Diseases: Examples of Drugs and Considerations for Future Integration in Clinical Practice." Curr Pharm Biotechnol 18(3): 231-241.

Chesney, E., G. M. Goodwin and S. Fazel (2014). "Risks of all-cause and suicide mortality in mental disorders: a meta-review." World Psychiatry 13(2): 153-160.

229

Choi, B. G., G. Vilahur, D. Yadegar, J. F. Viles-Gonzalez and J. J. Badimon (2006). "The role of high-density lipoprotein cholesterol in the prevention and possible treatment of cardiovascular diseases." Curr Mol Med 6(5): 571-587.

Choi, S. H., D. Ruggiero, R. Sorice, C. Song, T. Nutile, A. Vernon Smith, M. P. Concas, M. Traglia, C. Barbieri, N. C. Ndiaye, M. G. Stathopoulou, V. Lagou, G. B. Maestrale, C. Sala, S. Debette, P. Kovacs, L. Lind, J. Lamont, P. Fitzgerald, A. Tonjes, V. Gudnason, D. Toniolo, M. Pirastu, C. Bellenguez, R. S. Vasan, E. Ingelsson, A. L. Leutenegger, A. D. Johnson, A. L. DeStefano, S. Visvikis-Siest, S. Seshadri and M. Ciullo (2016). "Six Novel Loci Associated with Circulating VEGF Levels Identified by a Meta-analysis of Genome-Wide Association Studies." PLoS Genet 12(2): e1005874.

Clark-Raymond, A. and A. Halaris (2013). "VEGF and depression: a comprehensive assessment of clinical data." J Psychiatr Res 47(8): 1080-1087.

Clark-Raymond, A., E. Meresh, D. Hoppensteadt, J. Fareed, J. Sinacore and A. Halaris (2014). "Vascular Endothelial Growth Factor: a potential diagnostic biomarker for major depression." J Psychiatr Res 59: 22-27.

Colonna, V., T. Nutile, M. Astore, O. Guardiola, G. Antoniol, M. Ciullo and M. G. Persico (2007). "Campora: a young genetic isolate in South Italy." Hum Hered 64(2): 123-135.

Colonna, V., T. Nutile, R. R. Ferrucci, G. Fardella, M. Aversano, G. Barbujani and M. Ciullo (2009). "Comparing population structure as inferred from genealogical versus genetic information." Eur J Hum Genet 17(12): 1635-1641.

Colvin, P. L., E. Moriguchi, P. H. Barrett, J. S. Parks and L. L. Rudel (1999). "Small HDL particles containing two apoA-I molecules are precursors in vivo to medium and large HDL particles containing three and four apoA-I molecules in nonhuman primates." J Lipid Res 40(10): 1782-1792.

Combarros, O., M. Cortina-Borja, A. D. Smith and D. J. Lehmann (2009). "Epistasis in sporadic Alzheimer's disease." Neurobiol Aging 30(9): 1333-1349.

Conroy, R. M., K. Pyorala, A. P. Fitzgerald, S. Sans, A. Menotti, G. De Backer, D. De Bacquer, P. Ducimetiere, P. Jousilahti, U. Keil, I. Njolstad, R. G. Oganov, T. Thomsen, H. Tunstall-Pedoe, A. Tverdal, H. Wedel, P. Whincup, L. Wilhelmsen and I. M. Graham (2003). "Estimation of ten-year risk of fatal cardiovascular disease in Europe: the SCORE project." Eur Heart J 24(11): 987-1003.

230

Cooper, R., J. Cutler, P. Desvigne-Nickens, S. P. Fortmann, L. Friedman, R. Havlik, G. Hogelin, J. Marler, P. McGovern, G. Morosco, L. Mosca, T. Pearson, J. Stamler, D. Stryer and T. Thom (2000). "Trends and disparities in coronary heart disease, stroke, and other cardiovascular diseases in the United States: findings of the national conference on cardiovascular disease prevention." Circulation 102(25): 3137-3147.

Corder, E. H., A. M. Saunders, W. J. Strittmatter, D. E. Schmechel, P. C. Gaskell, G. W. Small, A. D. Roses, J. L. Haines and M. A. Pericak-Vance (1993). "Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families." Science 261(5123): 921-923.

D'Adamo, E., O. Guardamagna, F. Chiarelli, A. Bartuli, D. Liccardo, F. Ferrari and V. Nobili (2015). "Atherogenic dyslipidemia and cardiovascular risk factors in obese children." Int J Endocrinol 2015: 912047.

Davis, P. and J. Wagganer (2005). Lipid and Lipoprotein Metabolism. Lipid Metabolism and Health, CRC Press: 47-60.

De Ferrari, G. V., A. Papassotiropoulos, T. Biechele, F. Wavrant De-Vrieze, M. E. Avila, M. B. Major, A. Myers, K. Saez, J. P. Henriquez, A. Zhao, M. A. Wollmer, R. M. Nitsch, C. Hock, C. M. Morris, J. Hardy and R. T. Moon (2007). "Common genetic variation within the low-density lipoprotein receptor-related protein 6 and late-onset Alzheimer's disease." Proc Natl Acad Sci U S A 104(22): 9434-9439.

Deaton, A. M. and A. Bird (2011). "CpG islands and the regulation of transcription." Genes Dev 25(10): 1010-1022.

Debette, S., S. Visvikis-Siest, M. H. Chen, N. C. Ndiaye, C. Song, A. Destefano, R. Safa, M. Azimi Nezhad, D. Sawyer, J. B. Marteau, V. Xanthakis, G. Siest, L. Sullivan, M. Pfister, H. Smith, S. H. Choi, J. Lamont, L. Lind, Q. Yang, P. Fitzgerald, E. Ingelsson, R. S. Vasan and S. Seshadri (2011). "Identification of cis- and trans-acting genetic variants explaining up to half the variation in circulating vascular endothelial growth factor levels." Circ Res 109(5): 554-563.

Dorfman, R., Z. Khayat, T. Sieminowski, B. Golden and R. Lyons (2013). "Application of personalized medicine to chronic disease: a feasibility assessment." Clin Transl Med 2(1): 16.

Durrington, P. (2003). "Dyslipidaemia." Lancet 362(9385): 717-731.

Eckel, R. H., D. A. York, S. Rossner, V. Hubbard, I. Caterson, S. T. St Jeor, L. L. Hayman, R. M. Mullis and S. N. Blair (2004). "Prevention Conference VII: Obesity, a worldwide 231

epidemic related to heart disease and stroke: executive summary." Circulation 110(18): 2968-2975.

Etkin, A., C. Buchel and J. J. Gross (2015). "The neural bases of emotion regulation." Nat Rev Neurosci 16(11): 693-700.

Evans, D., F. U. Beil and J. Aberle (2013). "Resequencing the APOE gene reveals that rare mutations are not significant contributory factors in the development of type III hyperlipidemia." J Clin Lipidol 7(6): 671-674.

Evans, W. E. and M. V. Relling (2004). "Moving towards individualized medicine with pharmacogenomics." Nature 429(6990): 464-468.

Ference, B. A., H. N. Ginsberg, I. Graham, K. K. Ray, C. J. Packard, E. Bruckert, R. A. Hegele, R. M. Krauss, F. J. Raal, H. Schunkert, G. F. Watts, J. Boren, S. Fazio, J. D. Horton, L. Masana, S. J. Nicholls, B. G. Nordestgaard, B. van de Sluis, M. R. Taskinen, L. Tokgozoglu, U. Landmesser, U. Laufs, O. Wiklund, J. K. Stock, M. J. Chapman and A. L. Catapano (2017). "Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic, and clinical studies. A consensus statement from the European Atherosclerosis Society Consensus Panel." Eur Heart J.

Ferrara, N., H.-P. Gerber and J. LeCouter (2003). "The biology of VEGF and its receptors." Nat Med 9(6): 669-676.

Finegold, J. A., P. Asaria and D. P. Francis (2013). "Mortality from ischaemic heart disease by country, region, and age: statistics from World Health Organisation and United Nations." Int J Cardiol 168(2): 934-945.

Fischle, W., Y. Wang and C. D. Allis (2003). "Binary switches and modification cassettes in histone biology and beyond." Nature 425(6957): 475-479.

Froguel, P., N. C. Ndiaye, A. Bonnefond, N. Bouatia-Naji, A. Dechaume, G. Siest, B. Herbeth, M. Falchi, L. Bottolo, R. M. Gueant-Rodriguez, C. Lecoeur, M. R. Langlois, Y. Labrune, A. Ruokonen, S. El Shamieh, M. G. Stathopoulou, A. Morandi, C. Maffeis, D. Meyre, J. R. Delanghe, P. Jacobson, L. Sjostrom, L. M. Carlsson, A. Walley, P. Elliott, M. R. Jarvelin, G. V. Dedoussis and S. Visvikis-Siest (2012). "A genome-wide association study identifies rs2000999 as a strong genetic determinant of circulating haptoglobin levels." PLoS One 7(3): e32327.

GBD 2013 Mortality and Causes of Death Collaborators "Global, regional, and national age–sex specific all-cause and cause-specific mortality for 240

232

causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013." The Lancet 385(9963): 117-171.

Goldstein, J., H. Hobbs and M. Brown (2001). "The metabolic and molecular bases of inherited disease."

Grellier M, G. P., Fricain JC, Bidarra SJ, Renard M, Bareille R, Bourget C, Amédée J, Barbosa MA (2009). "The effect of the co-immobilization of human osteoprogenitors and endothelial cells within alginate microspheres on mineralization in a bone defect." Biomaterials 30: 3271-3278.

Grodstein , F., M. J. Stampfer , J. E. Manson , G. A. Colditz , W. C. Willett , B. Rosner , F. E. Speizer and C. H. Hennekens (1996). "Postmenopausal Estrogen and Progestin Use and the Risk of Cardiovascular Disease." New England Journal of Medicine 335(7): 453-461.

Grziwa, B., M. O. Grimm, C. L. Masters, K. Beyreuther, T. Hartmann and S. F. Lichtenthaler (2003). "The transmembrane domain of the amyloid precursor protein in microsomal membranes is on both sides shorter than predicted." J Biol Chem 278(9): 6803-6808.

Hachem, S. B. and A. D. Mooradian (2006). "Familial dyslipidaemias: an overview of genetics, pathophysiology and management." Drugs 66(15): 1949-1969.

Hardy, O. T., M. P. Czech and S. Corvera (2012). "What causes the insulin resistance underlying obesity?" Curr Opin Endocrinol Diabetes Obes 19(2): 81-87.

Hendrani, A. D., T. Adesiyun, R. Quispe, S. R. Jones, N. J. Stone, R. S. Blumenthal and S. S. Martin (2016). "Dyslipidemia management in primary prevention of cardiovascular disease: Current guidelines and strategies." World J Cardiol 8(2): 201-210.

Herzer, S., S. Meldner, H. J. Grone and V. Nordstrom (2015). "Fasting-Induced Lipolysis and Hypothalamic Insulin Signaling Are Regulated by Neuronal Glucosylceramide Synthase." Diabetes 64(10): 3363-3376.

Hixson, J. E. and D. T. Vernier (1990). "Restriction isotyping of human apolipoprotein E by gene amplification and cleavage with HhaI." J Lipid Res 31(3): 545-548.

Hokanson, J. E. and M. A. Austin (1996). "Plasma triglyceride level is a risk factor for cardiovascular disease independent of high-density lipoprotein cholesterol level: a

233

meta-analysis of population-based prospective studies." J Cardiovasc Risk 3(2): 213-219.

Huang, Y. and R. W. Mahley (2014). "Apolipoprotein E: structure and function in lipid metabolism, neurobiology, and Alzheimer's diseases." Neurobiol Dis 72 Pt A: 3-12.

Huang, Y. A., B. Zhou, M. Wernig and T. C. Sudhof (2017). "ApoE2, ApoE3, and ApoE4 Differentially Stimulate APP Transcription and Abeta Secretion." Cell 168(3): 427-441.e421.

Hurst, L. D. (2000). "Epistasis and the Evolutionary Process." Heredity 85(6): 625-626.

Hutchinson, L. (2013). "Biomarkers: New VEGF model-based biomarker." Nat Rev Clin Oncol 10(7): 366-366.

Huxley, R. R. and M. Woodward (2011). "Cigarette smoking as a risk factor for coronary heart disease in women compared with men: a systematic review and meta-analysis of prospective cohort studies." Lancet 378(9799): 1297-1305.

Hyde, C. L., M. W. Nagle, C. Tian, X. Chen, S. A. Paciga, J. R. Wendland, J. Y. Tung, D. A. Hinds, R. H. Perlis and A. R. Winslow (2016). "Identification of 15 genetic loci associated with risk of major depression in individuals of European descent." Nat Genet 48(9): 1031-1036.

Inoue, M., H. Itoh, M. Ueda, T. Naruko, A. Kojima, R. Komatsu, K. Doi, Y. Ogawa, N. Tamura, K. Takaya, T. Igaki, J. Yamashita, T. H. Chun, K. Masatsugu, A. E. Becker and K. Nakao (1998). "Vascular endothelial growth factor (VEGF) expression in human coronary atherosclerotic lesions: possible pathophysiological significance of VEGF in progression of atherosclerosis." Circulation 98(20): 2108-2116.

Isung, J., F. Mobarrez, P. Nordstrom, M. Asberg and J. Jokinen (2012). "Low plasma vascular endothelial growth factor (VEGF) associated with completed suicide." World J Biol Psychiatry 13(6): 468-473.

Jani, B. and C. Rajkumar (2006). "Ageing and vascular ageing." Postgraduate Medical Journal 82(968): 357-362.

Jin, K., Y. Zhu, Y. Sun, X. O. Mao, L. Xie and D. A. Greenberg (2002). "Vascular endothelial growth factor (VEGF) stimulates neurogenesis in vitro and in vivo." Proc Natl Acad Sci U S A 99(18): 11946-11950.

234

Johnell, O. and J. Kanis (2005). "Epidemiology of osteoporotic fractures." Osteoporos Int 16 Suppl 2: S3-7.

Jones, P. A. and S. B. Baylin (2007). "The epigenomics of cancer." Cell 128(4): 683-692.

Jousilahti, P., E. Vartiainen, J. Tuomilehto and P. Puska (1999). "Sex, Age, Cardiovascular Risk Factors, and Coronary Heart Disease." A Prospective Follow-Up Study of 14 786 Middle-Aged Men and Women in Finland 99(9): 1165-1172.

Kanis, J. A. (1994). "Assessment of fracture risk and its application to screening for postmenopausal osteoporosis: synopsis of a WHO report. WHO Study Group." Osteoporos Int 4(6): 368-381.

Kardia, S. L., S. M. Modell and P. A. Peyser (2003). "Family-centered approaches to understanding and preventing coronary heart disease." Am J Prev Med 24(2): 143-151.

Kato, M. and A. Serretti (2010). "Review and meta-analysis of antidepressant pharmacogenetic findings in major depressive disorder." Mol Psychiatry 15(5): 473-500.

Kern, P. A., M. Saghizadeh, J. M. Ong, R. J. Bosch, R. Deem and R. B. Simsolo (1995). "The expression of tumor necrosis factor in human adipose tissue. Regulation by obesity, weight loss, and relationship to lipoprotein lipase." J Clin Invest 95(5): 2111-2119.

Kessler, R. C. and E. J. Bromet (2013). "The epidemiology of depression across cultures." Annu Rev Public Health 34: 119-138.

Kim, J., J. M. Basak and D. M. Holtzman (2009). "The role of apolipoprotein E in Alzheimer's disease." Neuron 63(3): 287-303.

Kim, Y. J., K. S. Kim, S. Do, C. H. Kim, S. K. Kim and Y. C. Lee (1997). "Molecular cloning and expression of human alpha2,8-sialyltransferase (hST8Sia V)." Biochem Biophys Res Commun 235(2): 327-330.

King, M. C., J. H. Marks and J. B. Mandell (2003). "Breast and ovarian cancer risks due to inherited mutations in BRCA1 and BRCA2." Science 302(5645): 643-646.

Kivipelto, M., E. L. Helkala, M. P. Laakso, T. Hanninen, M. Hallikainen, K. Alhainen, S. Iivonen, A. Mannermaa, J. Tuomilehto, A. Nissinen and H. Soininen (2002).

235

"Apolipoprotein E epsilon4 allele, elevated midlife total cholesterol level, and high midlife systolic blood pressure are independent risk factors for late-life Alzheimer disease." Ann Intern Med 137(3): 149-155.

Klijs, B., S. Scholtens, J. J. Mandemakers, H. Snieder, R. P. Stolk and N. Smidt (2015). "Representativeness of the LifeLines Cohort Study." PLoS One 10(9): e0137203.

Koller, D. L., L. Liu, I. Alam, Q. Sun, M. J. Econs, T. Foroud and C. H. Turner (2008). "Epistatic effects contribute to variation in BMD in Fischer 344 x Lewis F2 rats." J Bone Miner Res 23(1): 41-47.

Kotan, Z., E. Sarandol, E. Kirhan, G. Ozkaya and S. Kirli (2012). "Serum brain-derived neurotrophic factor, vascular endothelial growth factor and leptin levels in patients with a diagnosis of severe major depressive disorder with melancholic features." Ther Adv Psychopharmacol 2(2): 65-74.

Kouzarides, T. (2007). "Chromatin Modifications and Their Function." Cell 128(4): 693-705.

Kupfer, D. J., E. Frank and M. L. Phillips (2012). "Major depressive disorder: new clinical, neurobiological, and treatment perspectives." Lancet 379(9820): 1045-1055.

Lee, B. H. and Y. K. Kim (2012). "Increased plasma VEGF levels in major depressive or manic episodes in patients with mood disorders." J Affect Disord 136(1-2): 181-184.

Li, M., C. D'Arcy and X. Meng (2016). "Maltreatment in childhood substantially increases the risk of adult depression and anxiety in prospective cohort studies: systematic review, meta-analysis, and proportional attributable fractions." Psychol Med 46(4): 717-730.

Liao, F., H. Yoon and J. Kim (2017). "Apolipoprotein E metabolism and functions in brain and its role in Alzheimer's disease." Curr Opin Lipidol 28(1): 60-67.

Libby, P. (2015). "Triglycerides on the rise: should we swap seats on the seesaw?" European Heart Journal 36(13): 774-776.

Licinio, J., C. Dong and M. L. Wong (2009). "Novel sequence variations in the brain-derived neurotrophic factor gene and association with major depression and antidepressant treatment response." Arch Gen Psychiatry 66(5): 488-497.

236

Liu Y, B. A., Jia S, et al (2012). "Intracellular VEGF regulates the balance between osteoblast and adipocyte differentiation." J Clin Invest 122: 3101-3113.

Lucotte, G., F. David, S. Visvikis, B. Leininger-Muller, G. Siest, M. C. Babron and R. Couderc (1993). "Apolipoprotein E-epsilon 4 allele and Alzheimer's disease." Lancet 342(8882): 1309.

Lüscher, T. F., U. Landmesser, A. von Eckardstein and A. M. Fogelman (2014). "High-Density Lipoprotein." Vascular Protective Effects, Dysfunction, and Potential as Therapeutic Target 114(1): 171-182.

Mahley, R. W. (2016). "Central Nervous System Lipoproteins: ApoE and Regulation of Cholesterol Metabolism." Arterioscler Thromb Vasc Biol 36(7): 1305-1315.

Mani, A., J. Radhakrishnan, H. Wang, A. Mani, M. A. Mani, C. Nelson-Williams, K. S. Carew, S. Mane, H. Najmabadi, D. Wu and R. P. Lifton (2007). "LRP6 mutation in a family with early coronary disease and metabolic risk factors." Science 315(5816): 1278-1282.

Mao, B., W. Wu, G. Davidson, J. Marhold, M. Li, B. M. Mechler, H. Delius, D. Hoppe, P. Stannek, C. Walter, A. Glinka and C. Niehrs (2002). "Kremen proteins are Dickkopf receptors that regulate Wnt/beta-catenin signalling." Nature 417(6889): 664-667.

Mayer, H., H. Bertram, W. Lindenmaier, T. Korff, H. Weber and H. Weich (2005). "Vascular endothelial growth factor (VEGF-A) expression in human mesenchymal stem cells: autocrine and paracrine role on osteoblastic and endothelial differentiation." J Cell Biochem 95(4): 827-839.

Mayer H, B. H., Lindenmaier W, Korff T, Weber H, Weich H (2005). "Vascular endothelial growth factor (VEGF-A) expression in human mesenchymal stem cells: autocrine and paracrine role on osteoblastic and endothelial differentiation." J Cell Biochem 95: 827-839.

Mayr-Wohlfart U, W. J., Hausser H, Kessler s, Günther KP, Dehio C, Puhl W, Brenner RE (2002). "Vascular endothelial growth factor stimulates chemotactic migration of primary human osteoblasts." Bone 30: 472-477.

Meadows, K. L. and H. I. Hurwitz (2012). "Anti-VEGF therapies in the clinic." Cold Spring Harb Perspect Med 2(10).

237

Mendelsohn , M. E. and R. H. Karas (1999). "The Protective Effects of Estrogen on the Cardiovascular System." New England Journal of Medicine 340(23): 1801-1811.

Mendis, S., P. Puska, B. Norrving, W. H. Organization, W. H. Federation and W. S. Organization (2011). Global Atlas on Cardiovascular Disease Prevention and Control, World Health Organization in collaboration with the World Heart Federation and the World Stroke Organization.

Midy V, P. J. (1994). "Vasculotropin/vascular endothelial growth factor induces differentiation in cultured osteoblasts." Biochem Biophys Res Commun 199: 380-386.

Miller, M., N. J. Stone, C. Ballantyne, V. Bittner, M. H. Criqui, H. N. Ginsberg, A. C. Goldberg, W. J. Howard, M. S. Jacobson, P. M. Kris-Etherton, T. A. Lennie, M. Levi, T. Mazzone and S. Pennathur (2011). "Triglycerides and Cardiovascular Disease." A Scientific Statement From the American Heart Association 123(20): 2292-2333.

Miller, S. A., D. D. Dykes and H. F. Polesky (1988). "A simple salting out procedure for extracting DNA from human nucleated cells." Nucleic Acids Res 16(3): 1215.

Moore, J. H. (2003). "The ubiquitous nature of epistasis in determining susceptibility to common human diseases." Hum Hered 56(1-3): 73-82.

Nabel , E. G. (2003). "Cardiovascular Disease." New England Journal of Medicine 349(1): 60-72.

Nagel, R. L. (2005). "Epistasis and the genetics of human diseases." C R Biol 328(7): 606-615.

National Heart Lung and Blood Institute task force (1998). "Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults--The Evidence Report. National Institutes of Health." Obes Res 6 Suppl 2: 51s-209s.

Neidhart, M. (2016). Preface. DNA Methylation and Complex Human Disease. Oxford, Academic Press: xiii.

NIH. (2015). "Handout on Health: Osteoporosis." from https://www.bones.nih.gov/health-info/bone/osteoporosis/osteoporosis-hoh.

Noel, P. H., J. W. Williams, Jr., J. Unutzer, J. Worchel, S. Lee, J. Cornell, W. Katon, L. H. Harpole and E. Hunkeler (2004). "Depression and comorbid illness in elderly primary

238

care patients: impact on multiple domains of health status and well-being." Ann Fam Med 2(6): 555-562.

Nolte, I. M. and P. J. van der Most (2017). "Missing heritability: is the gap closing? An analysis of 32 complex traits in the Lifelines Cohort Study." 25(7): 877-885.

Nordestgaard, B. G., M. J. Chapman, S. E. Humphries, H. N. Ginsberg, L. Masana, O. S. Descamps, O. Wiklund, R. A. Hegele, F. J. Raal, J. C. Defesche, A. Wiegman, R. D. Santos, G. F. Watts, K. G. Parhofer, G. K. Hovingh, P. T. Kovanen, C. Boileau, M. Averna, J. Borén, E. Bruckert, A. L. Catapano, J. A. Kuivenhoven, P. Pajukanta, K. Ray, A. F. H. Stalenhoef, E. Stroes, M.-R. Taskinen, A. Tybjærg-Hansen and P. for the European Atherosclerosis Society Consensus (2013). "Familial hypercholesterolaemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease : Consensus Statement of the European Atherosclerosis Society." European Heart Journal 34(45): 3478-3490.

Notkola, I. L., R. Sulkava, J. Pekkanen, T. Erkinjuntti, C. Ehnholm, P. Kivinen, J. Tuomilehto and A. Nissinen (1998). "Serum total cholesterol, apolipoprotein E epsilon 4 allele, and Alzheimer's disease." Neuroepidemiology 17(1): 14-20.

Nowak, D. G., E. M. Amin, E. S. Rennel, C. Hoareau-Aveilla, M. Gammons, G. Damodoran, M. Hagiwara, S. J. Harper, J. Woolard, M. R. Ladomery and D. O. Bates (2010). "Regulation of vascular endothelial growth factor (VEGF) splicing from pro-angiogenic to anti-angiogenic isoforms: a novel therapeutic strategy for angiogenesis." J Biol Chem 285(8): 5532-5540.

O'Donnell, C. J. and E. G. Nabel (2008). "Cardiovascular Genomics, Personalized Medicine, and the National Heart, Lung, and Blood Institute." Part I: The Beginning of an Era 1(1): 51-57.

O'Rourke, M. F., M. E. Safar and V. Dzau (2010). "The Cardiovascular Continuum extended: aging effects on the aorta and microvasculature." Vasc Med 15(6): 461-468.

Okano, M., D. W. Bell, D. A. Haber and E. Li (1999). "DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development." Cell 99(3): 247-257.

Olsson, A.-K., A. Dimberg, J. Kreuger and L. Claesson-Welsh (2006). "VEGF receptor signalling ? in control of vascular function." Nat Rev Mol Cell Biol 7(5): 359-371.

239

Olsson, A. K., A. Dimberg, J. Kreuger and L. Claesson-Welsh (2006). "VEGF receptor signalling - in control of vascular function." Nat Rev Mol Cell Biol 7(5): 359-371.

Penninx, B. W., Y. Milaneschi, F. Lamers and N. Vogelzangs (2013). "Understanding the somatic consequences of depression: biological mechanisms and the role of depression symptom profile." BMC Med 11: 129.

Pinherio, J. C. and D. M. Bates. (2004). Mixed-Effects Models in S and S-PLUS, Statistics and Computing Series, Springer.

Poirier, P. and R. H. Eckel (2002). "Obesity and cardiovascular disease." Curr Atheroscler Rep 4(6): 448-453.

Poole, S., D. Walker, R. E. Gaines Das, J. R. Gallimore and M. B. Pepys (1998). "The first international standard for serum amyloid A protein (SAA). Evaluation in an international collaborative study." J Immunol Methods 214(1-2): 1-10.

Qiu, J. (2006). "Epigenetics: Unfinished symphony." Nature 441(7090): 143-145.

Rashid, P., J. Leonardi-Bee and P. Bath (2003). "Blood Pressure Reduction and Secondary Prevention of Stroke and Other Vascular Events." A Systematic Review 34(11): 2741-2748.

Rebeck, G. W., M. J. LaDu, S. Estus, G. Bu and E. J. Weeber (2006). "The generation and function of soluble apoE receptors in the CNS." Mol Neurodegener 1: 15.

Regitz-Zagrosek, V., S. Oertelt-Prigione, E. Prescott, F. Franconi, E. Gerdts, A. Foryst-Ludwig, A. H. E. M. Maas, A. Kautzky-Willer, D. Knappe-Wegner, U. Kintscher, K. H. Ladwig, K. Schenck-Gustafsson and V. Stangl (2016). "Gender in cardiovascular diseases: impact on clinical manifestations, management, and outcomes." European Heart Journal 37(1): 24-34.

Reitz, C. (2013). "Dyslipidemia and the risk of Alzheimer's disease." Curr Atheroscler Rep 15(3): 307.

Rodriguez-Antona, C. and M. Ingelman-Sundberg (2006). "Cytochrome P450 pharmacogenetics and cancer." Oncogene 25(11): 1679-1691.

Sacks, F. M., M. A. Pfeffer, L. A. Moye, J. L. Rouleau, J. D. Rutherford, T. G. Cole, L. Brown, J. W. Warnica, J. M. Arnold, C. C. Wun, B. R. Davis and E. Braunwald (1996). "The effect of pravastatin on coronary events after myocardial infarction in patients

240

with average cholesterol levels. Cholesterol and Recurrent Events Trial investigators." N Engl J Med 335(14): 1001-1009.

Sarzani, R., F. Salvi, M. Bordicchia, F. Guerra, I. Battistoni, G. Pagliariccio, L. Carbonari, P. Dessi-Fulgheri and A. Rappelli (2011). "Carotid artery atherosclerosis in hypertensive patients with a functional LDL receptor-related protein 6 gene variant." Nutr Metab Cardiovasc Dis 21(2): 150-156.

Scheltens, P., K. Blennow, M. M. Breteler, B. de Strooper, G. B. Frisoni, S. Salloway and W. M. Van der Flier (2016). "Alzheimer's disease." Lancet 388(10043): 505-517.

Schiele, F., D. De Bacquer, M. Vincent-Viry, U. Beisiegel, C. Ehnholm, A. Evans, A. Kafatos, M. C. Martins, S. Sans, C. Sass, S. Visvikis, G. De Backer and G. Siest (2000). "Apolipoprotein E serum concentration and polymorphism in six European countries: the ApoEurope Project." Atherosclerosis 152(2): 475-488.

Scholtens, S., N. Smidt, M. A. Swertz, S. J. Bakker, A. Dotinga, J. M. Vonk, F. van Dijk, S. K. van Zon, C. Wijmenga, B. H. Wolffenbuttel and R. P. Stolk (2015). "Cohort Profile: LifeLines, a three-generation cohort study and biobank." Int J Epidemiol 44(4): 1172-1180.

Serretti, A., M. Kato, D. De Ronchi and T. Kinoshita (2007). "Meta-analysis of serotonin transporter gene promoter polymorphism (5-HTTLPR) association with selective serotonin reuptake inhibitor efficacy in depressed patients." Mol Psychiatry 12(3): 247-257.

Siest, G., P. Bertrand, B. Herbeth, M. Vincent-Viry, F. Schiele, C. Sass and S. Visvikis (2000). "Apolipoprotein E polymorphisms and concentration in chronic diseases and drug responses." Clin Chem Lab Med 38(9): 841-852.

Siest, G., P. Bertrand, B. Qin, B. Herbeth, J. M. Serot, L. Masana, J. Ribalta, A. P. Passmore, A. Evans, M. Ferrari, M. Franceschi, J. Shepherd, M. Cuchel, U. Beisiegel, K. Zuchowsky, A. S. Rukavina, J. Sertic, M. Stojanov, V. Kostic, A. Mitrevski, V. Petrova, C. Sass, A. Merched, J. T. Salonen, L. Tiret and S. Visvikis (2000). "Apolipoprotein E polymorphism and serum concentration in Alzheimer's disease in nine European centres: the ApoEurope study. ApoEurope group." Clin Chem Lab Med 38(8): 721-730.

Siris, E. S., R. Adler, J. Bilezikian, M. Bolognese, B. Dawson-Hughes, M. J. Favus, S. T. Harris, S. M. Jan de Beur, S. Khosla, N. E. Lane, R. Lindsay, A. D. Nana, E. S. Orwoll, K. Saag, S. Silverman and N. B. Watts (2014). "The clinical diagnosis of osteoporosis: a

241

position statement from the National Bone Health Alliance Working Group." Osteoporosis International 25(5): 1439-1443.

Solberg, L. A. and J. P. Strong (1983). "Risk factors and atherosclerotic lesions. A review of autopsy studies." Arteriosclerosis, Thrombosis, and Vascular Biology 3(3): 187-198.

Solomon, A., M. Kivipelto, B. Wolozin, J. Zhou and R. A. Whitmer (2009). "Midlife serum cholesterol and increased risk of Alzheimer's and vascular dementia three decades later." Dement Geriatr Cogn Disord 28(1): 75-80.

Stathopoulou, M. G., A. Bonnefond, N. C. Ndiaye, M. Azimi-Nezhad, S. El Shamieh, A. Saleh, M. Rancier, G. Siest, J. Lamont, P. Fitzgerald and S. Visvikis-Siest (2013). "A common variant highly associated with plasma VEGFA levels also contributes to the variation of both LDL-C and HDL-C." J Lipid Res 54(2): 535-541.

Stein, E. A. and G. L. Myers (1995). "National Cholesterol Education Program recommendations for triglyceride measurement: executive summary. The National Cholesterol Education Program Working Group on Lipoprotein Measurement." Clin Chem 41(10): 1421-1426.

Stenger, C., C. Corbier and F. T. Yen (2012). "Structure and Function of the Lipolysis Stimulated Lipoprotein Receptor In Chemical Biology, Prof. Deniz Ekinci (Ed.)." InTech: pp. 267-292.

Stenger, C., A. Pincon, M. Hanse, L. Royer, A. Comte, V. Koziel, J. L. Olivier, T. Pillot and F. T. Yen (2012). "Brain region-specific immunolocalization of the lipolysis-stimulated lipoprotein receptor (LSR) and altered cholesterol distribution in aged LSR+/- mice." J Neurochem 123(4): 467-476.

Stephen J. McPhee (2012). Current medical diagnosis & treatment 2012, New York: McGraw-Hill Medical.

Stevens, M., J. B. Cheng, D. Li, M. Xie, C. Hong, C. L. Maire, K. L. Ligon, M. Hirst, M. A. Marra, J. F. Costello and T. Wang (2013). "Estimating absolute methylation levels at single-CpG resolution from methylation enrichment and restriction enzyme sequencing methods." Genome Res 23(9): 1541-1553.

Stolk, R. P., J. G. Rosmalen, D. S. Postma, R. A. de Boer, G. Navis, J. P. Slaets, J. Ormel and B. H. Wolffenbuttel (2008). "Universal risk factors for multifactorial diseases: LifeLines: a three-generation population-based study." Eur J Epidemiol 23(1): 67-74.

242

Street J, L. B. (2009). "Vascular endothelial growth factor regulates osteoblast survival - evidence for an autocrine feedback mechanism." J Orthop Surg Res 16: 4-19.

Suto, K., Y. Yamazaki, T. Morita and H. Mizuno (2005). "Crystal structures of novel vascular endothelial growth factors (VEGF) from snake venoms: insight into selective VEGF binding to kinase insert domain-containing receptor but not to fms-like tyrosine kinase-1." J Biol Chem 280(3): 2126-2131.

Takahashi, H. and M. Shibuya (2005). "The vascular endothelial growth factor (VEGF)/VEGF receptor system and its role under physiological and pathological conditions." Clin Sci (Lond) 109(3): 227-241.

Tan YY, Y. Y., Chai L, Wong RW, Rabie AB (2010). "Effects of vascular endothelial growth factor (VEGF] on MC3T3-E1 " Orthod Craniofac Res 13: 223-228.

Teixeira, P. J., L. B. Sardinha, S. B. Going and T. G. Lohman (2001). "Total and regional fat and serum cardiovascular disease risk factors in lean and obese children and adolescents." Obes Res 9(8): 432-442.

Thornton-Wells, T. A., J. H. Moore and J. L. Haines (2004). "Genetics, statistics and human disease: analytical retooling for complexity." Trends Genet 20(12): 640-647.

Tischer, E., R. Mitchell, T. Hartman, M. Silva, D. Gospodarowicz, J. C. Fiddes and J. A. Abraham (1991). "The human gene for vascular endothelial growth factor. Multiple protein forms are encoded through alternative exon splicing." J Biol Chem 266(18): 11947-11954.

Toth, P. P. 2005. "The Good Cholesterol." High-Density Lipoprotein 111(5): e89-e91.

Toth, P. P., P. J. Barter, R. S. Rosenson, W. E. Boden, M. J. Chapman, M. Cuchel, R. B. D'Agostino, Sr., M. H. Davidson, W. S. Davidson, J. W. Heinecke, R. H. Karas, A. Kontush, R. M. Krauss, M. Miller and D. J. Rader (2013). "High-density lipoproteins: a consensus statement from the National Lipid Association." J Clin Lipidol 7(5): 484-525.

Trueta J, B. A. (1963). "The vascular contribution to osteogenesis. V. The vasculature supplying the epiphysial cartilage in rachitic rats." Journal of Bone and Joint Surgery British 45: 572-581.

243

Upadhyay, R. K. (2015). "Emerging Risk Biomarkers in Cardiovascular Diseases and Disorders." Journal of Lipids 2015: 50.

Varbo, A., M. Benn, A. Tybjaerg-Hansen, A. B. Jorgensen, R. Frikke-Schmidt and B. G. Nordestgaard (2013). "Remnant cholesterol as a causal risk factor for ischemic heart disease." J Am Coll Cardiol 61(4): 427-436.

Visvikis-Siest, S. and G. Siest (2008). "The STANISLAS Cohort: a 10-year follow-up of supposed healthy families. Gene-environment interactions, reference values and evaluation of biomarkers in prevention of cardiovascular diseases." Clin Chem Lab Med 46(6): 733-747.

Vos, T., R. M. Barber, B. Bell, A. Bertozzi-Villa, S. Biryukov, I. Bolliger, F. Charlson, A. Davis, L. Degenhardt, D. Dicker, L. Duan, H. Erskine, V. L. Feigin, A. J. Ferrari, C. Fitzmaurice, T. Fleming, N. Graetz, C. Guinovart, J. Haagsma, G. M. Hansen, S. W. Hanson, K. R. Heuton, H. Higashi, N. Kassebaum, H. Kyu, E. Laurie, X. Liang, K. Lofgren, R. Lozano, M. F. MacIntyre, M. Moradi-Lakeh, M. Naghavi, G. Nguyen, S. Odell, K. Ortblad, D. A. Roberts, G. A. Roth, L. Sandar, P. T. Serina, J. D. Stanaway, C. Steiner, B. Thomas, S. E. Vollset, H. Whiteford, T. M. Wolock, P. Ye, M. Zhou, M. A. Ãvila, G. M. Aasvang, C. Abbafati, A. A. Ozgoren, F. Abd-Allah, M. I. A. Aziz, S. F. Abera, V. Aboyans, J. P. Abraham, B. Abraham, I. Abubakar, L. J. Abu-Raddad, N. M. E. Abu-Rmeileh, T. C. Aburto, T. Achoki, I. N. Ackerman, A. Adelekan, Z. Ademi, A. K. Adou, J. C. Adsuar, J. Arnlov, E. E. Agardh, M. J. Al Khabouri, S. S. Alam, D. Alasfoor, M. I. Albittar, M. A. Alegretti, A. V. Aleman, Z. A. Alemu, R. Alfonso-Cristancho, S. Alhabib, R. Ali, F. Alla, P. Allebeck, P. J. Allen, M. A. AlMazroa, U. Alsharif, E. Alvarez, N. Alvis-Guzman, O. Ameli, H. Amini, W. Ammar, B. O. Anderson, H. R. Anderson, C. A. T. Antonio, P. Anwari, H. Apfel, V. S. A. Arsenijevic, A. Artaman, R. J. Asghar, R. Assadi, L. S. Atkins, C. Atkinson, A. Badawi, M. C. Bahit, T. Bakfalouni, K. Balakrishnan, S. Balalla, A. Banerjee, S. L. Barker-Collo, S. Barquera, L. Barregard, L. H. Barrero, S. Basu, A. Basu, A. Baxter, J. Beardsley, N. Bedi, E. Beghi, T. Bekele, M. L. Bell, C. Benjet, D. A. Bennett, I. M. Bensenor, H. Benzian, E. Bernabe, T. J. Beyene, N. Bhala, A. Bhalla, Z. Bhutta, K. Bienhoff, B. Bikbov, A. B. Abdulhak, J. D. Blore, F. M. Blyth, M. A. Bohensky, B. B. Basara, G. Borges, N. M. Bornstein, D. Bose, S. Boufous, R. R. Bourne, L. N. Boyers, M. Brainin, M. Brauer, C. E. G. Brayne, A. Brazinova, N. J. K. Breitborde, H. Brenner, A. D. M. Briggs, P. M. Brooks, J. Brown, T. S. Brugha, R. Buchbinder, G. C. Buckle, G. Bukhman, A. G. Bulloch, M. Burch, R. Burnett, R. Cardenas, N. L. Cabral, I. R. C. Nonato, J. C. Campuzano, J. R. Carapetis, D. O. Carpenter, V. Caso, C. A. Castaneda-Orjuela, F. Catala-Lopez, V. K. Chadha, J.-C. Chang, H. Chen, W. Chen, P. P. Chiang, O. Chimed-Ochir, R. Chowdhury, H. Christensen, C. A. Christophi, S. S. Chugh, M. Cirillo, M. Coggeshall, A. Cohen, V. Colistro, S. M. Colquhoun, A. G. Contreras, L. T. Cooper, C. Cooper, K. Cooperrider, J. Coresh, M. Cortinovis, M. H. Criqui, J. A. Crump, L. Cuevas-Nasu, R. Dandona, L. Dandona, E. Dansereau, H. G. Dantes, P. I. Dargan, G. Davey, D. V. Davitoiu, A. Dayama, V. De la Cruz-Gongora, S. F. de la Vega, D. De Leo, B. 244

del Pozo-Cruz, R. P. Dellavalle, K. Deribe, S. Derrett, D. C. Des Jarlais, M. Dessalegn, G. A. deVeber, S. D. Dharmaratne, C. Diaz-Torne, E. L. Ding, K. Dokova, E. R. Dorsey, T. R. Driscoll, H. Duber, A. M. Durrani, K. M. Edmond, R. G. Ellenbogen, M. Endres, S. P. Ermakov, B. Eshrati, A. Esteghamati, K. Estep, S. Fahimi, F. Farzadfar, D. F. J. Fay, D. T. Felson, S.-M. Fereshtehnejad, J. G. Fernandes, C. P. Ferri, A. Flaxman, N. Foigt, K. J. Foreman, F. G. R. Fowkes, R. C. Franklin, T. Furst, N. D. Futran, B. J. Gabbe, F. G. Gankpe, F. A. Garcia-Guerra, J. M. Geleijnse, B. D. Gessner, K. B. Gibney, R. F. Gillum, I. A. Ginawi, M. Giroud, G. Giussani, S. Goenka, K. Goginashvili, P. Gona, T. G. de Cosio, R. A. Gosselin, C. C. Gotay, A. Goto, H. N. Gouda, R. l. Guerrant, H. C. Gugnani, D. Gunnell, R. Gupta, R. Gupta, R. A. Gutierrez, N. Hafezi-Nejad, H. Hagan, Y. Halasa, R. R. Hamadeh, H. Hamavid, M. Hammami, G. J. Hankey, Y. Hao, H. L. Harb, J. M. Haro, R. Havmoeller, R. J. Hay, S. Hay, M. T. Hedayati, I. B. H. Pi, P. Heydarpour, M. Hijar, H. W. Hoek, H. J. Hoffman, J. C. Hornberger, H. D. Hosgood, M. Hossain, P. J. Hotez, D. G. Hoy, M. Hsairi, H. Hu, G. Hu, J. J. Huang, C. Huang, L. Huiart, A. Husseini, M. Iannarone, K. M. Iburg, K. Innos, M. Inoue, K. H. Jacobsen, S. K. Jassal, P. Jeemon, P. N. Jensen, V. Jha, G. Jiang, Y. Jiang, J. B. Jonas, J. Joseph, K. Juel, H. Kan, A. Karch, C. Karimkhani, G. Karthikeyan, R. Katz, A. Kaul, N. Kawakami, D. S. Kazi, A. H. Kemp, A. P. Kengne, Y. S. Khader, S. E. A. H. Khalifa, E. A. Khan, G. Khan, Y.-H. Khang, I. Khonelidze, C. Kieling, D. Kim, S. Kim, R. W. Kimokoti, Y. Kinfu, J. M. Kinge, B. M. Kissela, M. Kivipelto, L. Knibbs, A. K. Knudsen, Y. Kokubo, S. Kosen, A. Kramer, M. Kravchenko, R. V. Krishnamurthi, S. Krishnaswami, B. K. Defo, B. K. Bicer, E. J. Kuipers, V. S. Kulkarni, K. Kumar, G. A. Kumar, G. F. Kwan, T. Lai, R. Lalloo, H. Lam, Q. Lan, V. C. Lansingh, H. Larson, A. Larsson, A. E. B. Lawrynowicz, J. L. Leasher, J.-T. Lee, J. Leigh, R. Leung, M. Levi, B. Li, Y. Li, Y. Li, J. liang, S. Lim, H.-H. Lin, M. Lind, M. P. Lindsay, S. E. Lipshultz, S. Liu, B. K. Lloyd, S. L. Ohno, G. Logroscino, K. J. Looker, A. D. Lopez, N. Lopez-Olmedo, J. Lortet-Tieulent, P. A. Lotufo, N. Low, R. M. Lucas, R. Lunevicius, R. A. Lyons, J. Ma, S. Ma, M. T. Mackay, M. Majdan, R. Malekzadeh, C. C. Mapoma, W. Marcenes, L. M. March, C. Margono, G. B. Marks, M. B. Marzan, J. R. Masci, A. J. Mason-Jones, R. G. Matzopoulos, B. M. Mayosi, T. T. Mazorodze, N. W. McGill, J. J. McGrath, M. McKee, A. McLain, B. J. McMahon, P. A. Meaney, M. M. Mehndiratta, F. Mejia-Rodriguez, W. Mekonnen, Y. A. Melaku, M. Meltzer, Z. A. Memish, G. Mensah, A. Meretoja, F. A. Mhimbira, R. Micha, T. R. Miller, E. J. Mills, P. B. Mitchell, C. N. Mock, T. E. Moffitt, N. M. Ibrahim, K. A. Mohammad, A. H. Mokdad, G. L. Mola, L. Monasta, M. Montico, T. J. Montine, A. R. Moore, A. E. Moran, L. Morawska, R. Mori, J. Moschandreas, W. N. Moturi, M. Moyer, D. Mozaffarian, U. O. Mueller, M. Mukaigawara, M. E. Murdoch, J. Murray, K. S. Murthy, P. Naghavi, Z. Nahas, A. Naheed, K. S. Naidoo, L. Naldi, D. Nand, V. Nangia, K. M. V. Narayan, D. Nash, C. Nejjari, S. P. Neupane, L. M. Newman, C. R. Newton, M. Ng, F. N. Ngalesoni, N. T. Nhung, M. I. Nisar, S. Nolte, O. F. Norheim, R. E. Norman, B. Norrving, L. Nyakarahuka, I. H. Oh, T. Ohkubo, S. B. Omer, J. N. Opio, A. Ortiz, J. D. Pandian, C. I. A. Panelo, C. Papachristou, E.-K. Park, C. D. Parry, A. J. P. Caicedo, S. B. Patten, V. K. Paul, B. I. Pavlin, N. Pearce, L. S. Pedraza, C. A. Pellegrini, D. M. Pereira, F. P. Perez-Ruiz, N. Perico, A. Pervaiz, K. Pesudovs, C. B. Peterson, M. Petzold, M. R. Phillips, D. Phillips, B. Phillips, F. B. Piel, D. Plass, D. Poenaru, G. V.

245

Polanczyk, S. Polinder, C. A. Pope, S. Popova, R. G. Poulton, F. Pourmalek, D. Prabhakaran, N. M. Prasad, D. Qato, D. A. Quistberg, A. Rafay, K. Rahimi, V. Rahimi-Movaghar, S. u. Rahman, M. Raju, I. Rakovac, S. M. Rana, H. Razavi, A. Refaat, J. Rehm, G. Remuzzi, S. Resnikoff, A. L. Ribeiro, P. M. Riccio, L. Richardson, J. H. Richardus, A. M. Riederer, M. Robinson, A. Roca, A. Rodriguez, D. Rojas-Rueda, L. Ronfani, D. Rothenbacher, N. Roy, G. M. Ruhago, N. Sabin, R. L. Sacco, K. Ksoreide, S. Saha, R. Sahathevan, M. A. Sahraian, U. Sampson, J. R. Sanabria, L. Sanchez-Riera, I. S. Santos, M. Satpathy, J. E. Saunders, M. Sawhney, M. I. Saylan, P. Scarborough, B. Schoettker, I. J. C. Schneider, D. C. Schwebel, J. G. Scott, S. Seedat, S. G. Sepanlou, B. Serdar, E. E. Servan-Mori, K. Shackelford, A. Shaheen, S. Shahraz, T. S. Levy, S. Shangguan, J. She, S. Sheikhbahaei, D. S. Shepard, P. Shi, K. Shibuya, Y. Shinohara, R. Shiri, K. Shishani, I. Shiue, M. G. Shrime, I. D. Sigfusdottir, D. H. Silberberg, E. P. Simard, S. Sindi, J. A. Singh, L. Singh, V. Skirbekk, K. Sliwa, M. Soljak, S. Soneji, S. S. Soshnikov, P. Speyer, L. A. Sposato, C. T. Sreeramareddy, H. Stoeckl, V. K. Stathopoulou, N. Steckling, M. B. Stein, D. J. Stein, T. J. Steiner, A. Stewart, E. Stork, L. J. Stovner, K. Stroumpoulis, L. Sturua, B. F. Sunguya, M. Swaroop, B. L. Sykes, K. M. Tabb, K. Takahashi, F. Tan, N. Tandon, D. Tanne, M. Tanner, M. Tavakkoli, H. R. Taylor, B. J. Te Ao, A. M. Temesgen, M. T. Have, E. Y. Tenkorang, A. S. Terkawi, A. M. Theadom, E. Thomas, A. L. Thorne-Lyman, A. G. Thrift, I. M. Tleyjeh, M. Tonelli, F. Topouzis, J. A. Towbin, H. Toyoshima, J. Traebert, B. X. Tran, L. Trasande, M. Trillini, T. Truelsen, U. Trujillo, M. Tsilimbaris, E. M. Tuzcu, K. N. Ukwaja, E. A. Undurraga, S. B. Uzun, W. H. van Brakel, S. van de Vijver, R. V. Dingenen, C. H. van Gool, Y. Y. Varakin, T. J. Vasankari, M. S. Vavilala, L. J. Veerman, G. Velasquez-Melendez, N. Venketasubramanian, L. Vijayakumar, S. Villalpando, F. S. Violante, V. V. Vlassov, S. Waller, M. T. Wallin, X. Wan, L. Wang, J. Wang, Y. Wang, T. S. Warouw, S. Weichenthal, E. Weiderpass, R. G. Weintraub, A. Werdecker, K. R. R. Wessells, R. Westerman, J. D. Wilkinson, H. C. Williams, T. N. Williams, S. M. Woldeyohannes, C. D. A. Wolfe, J. Q. Wong, H. Wong, A. D. Woolf, J. L. Wright, B. Wurtz, G. Xu, G. Yang, Y. Yano, M. A. Yenesew, G. K. Yentur, P. Yip, N. Yonemoto, S.-J. Yoon, M. Younis, C. Yu, K. Y. Kim, M. E. S. Zaki, Y. Zhang, Z. Zhao, Y. Zhao, J. Zhu, D. Zonies, J. R. Zunt, J. A. Salomon and C. J. L. Murray (2015). "Global, regional, and national incidence, prevalence, and years lived with disability for 301 acute and chronic diseases and injuries in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013." The Lancet 386(9995): 743-800.

Waddington, C. H. (2012). "The epigenotype. 1942." Int J Epidemiol 41(1): 10-13.

Warnick, G. R. and P. D. Wood (1995). "National Cholesterol Education Program recommendations for measurement of high-density lipoprotein cholesterol: executive summary. The National Cholesterol Education Program Working Group on Lipoprotein Measurement." Clin Chem 41(10): 1427-1433.

246

WHO (2014). Global status report on non-communicable diseases 2014. . Geneva, World Health Organization.

Whooley, M. A. and J. M. Wong (2013). "Depression and cardiovascular disorders." Annu Rev Clin Psychol 9: 327-354.

Wiltshire, S., J. T. Bell, C. J. Groves, C. Dina, A. T. Hattersley, T. M. Frayling, M. Walker, G. A. Hitman, M. Vaxillaire, M. Farrall, P. Froguel and M. I. McCarthy (2006). "Epistasis between type 2 diabetes susceptibility Loci on chromosomes 1q21-25 and 10q23-26 in northern Europeans." Ann Hum Genet 70(Pt 6): 726-737.

Woolard, J., H. S. Bevan, S. J. Harper and D. O. Bates (2009). "Molecular diversity of VEGF-A as a regulator of its biological activity." Microcirculation 16(7): 572-592.

World Health Organization (2016). "Cardiovascular diseases (CVDs)." WHO(http://www.who.int/mediacentre/factsheets/fs317/en/).

World Heart Federation. "Risk factors." from https://www.world-heart-federation.org/resources/risk-factors/.

Wu, C. and J. R. Morris (2001). "Genes, genetics, and epigenetics: a correspondence." Science 293(5532): 1103-1105.

Xiong, H., D. Callaghan, A. Jones, D. G. Walker, L. F. Lue, T. G. Beach, L. I. Sue, J. Woulfe, H. Xu, D. B. Stanimirovic and W. Zhang (2008). "Cholesterol retention in Alzheimer's brain is responsible for high beta- and gamma-secretase activities and Abeta production." Neurobiol Dis 29(3): 422-437.

Yancopoulos, G. D., S. Davis, N. W. Gale, J. S. Rudge, S. J. Wiegand and J. Holash (2000). "Vascular-specific growth factors and blood vessel formation." Nature 407(6801): 242-248.

Yasuhara, T., T. Shingo and I. Date (2004). "The potential role of vascular endothelial growth factor in the central nervous system." Rev Neurosci 15(4): 293-307.

Yen, F. T., C. J. Mann, L. M. Guermani, N. F. Hannouche, N. Hubert, C. A. Hornick, V. N. Bordeau, G. Agnani and B. E. Bihain (1994). "Identification of a lipolysis-stimulated receptor that is distinct from the LDL receptor and the LDL receptor-related protein." Biochemistry 33(5): 1172-1180.

247

Yen, F. T., O. Roitel, L. Bonnard, V. Notet, D. Pratte, C. Stenger, E. Magueur and B. E. Bihain (2008). "Lipolysis stimulated lipoprotein receptor: a novel molecular link between hyperlipidemia, weight gain, and atherosclerosis in mice." J Biol Chem 283(37): 25650-25659.

Zhang, J. and Q. Liu (2015). "Cholesterol metabolism and homeostasis in the brain." Protein Cell 6(4): 254-264.

Zhao, R. (2012). "Immune regulation of osteoclast function in postmenopausal osteoporosis: a critical interdisciplinary perspective." Int J Med Sci 9(9): 825-832.

248

Abstract Chronic diseases, like cardiovascular diseases (CVD), Alzheimer’s disease (AD), depression and osteoporosis, are major causes of mortality in the world. Identification of common to those diseases risk factors could help for a better-monitored ‘healthy’ aging, by promotion of personalised strategies for risk prediction, early prevention and adequate treatment, all taking into account the very often existing comorbidities. In this thesis, 8 publications have been developed. Initially, in a review paper, I have summarised the current challenges and opportunities of pharmacogenomics of CVD medications. I have participated in the formation of an international consortium, the VEGF Consortium, and I have participated in a study that identified significant epistatic interactions between polymorphisms that regulate the levels of VEGF and their effects on blood pressure and adiposity indexes. I have also demonstrated that one genetic marker of VEGF, rs4416670, was significantly associated with an increased risk for depression. Furthermore, I have reported two significant interactions between VEGF-related variants affecting the femoral neck bone mineral density in post-menopausal women. I have focused also on two markers linked with lipids metabolism: the apolipoprotein E (APOE) and the lipolysis-stimulated receptor (LSR). I have found that the LSR variant rs916147 can interact with APOE in a way that reverses the protective effect of the ε2 allele of APOE on blood lipids, thus providing new insights in the mechanisms underlying type III hyperlipoproteinemia. Epistatic interactions between these two genes have also been shown to increase the risk of AD, even in the absence of the known risk allele APOE ε4. Finally, I have performed epigenome-wide association studies (EWAS) on central obesity and blood lipid traits in healthy individuals. The results suggest that one methylation probe could affect waist circumference through an insulin-signaling pathway. Furthermore, two methylations probes were associated with triglycerides levels through genes linked with genetic heart diseases (PRKAG2) and with inhibition of the Wnt/beta-catenin signaling that is involved in CVD and AD development (KREMEN2). In conclusion, this thesis used the study of epistasis and epigenetics and identified complex inter-relationships between VEGF, LSR, APOE and different chronic diseases (CVD, AD, osteoporosis, depression) and novel mechanisms that link disease development with DNA methylation, thus demonstrating their role as common denominators of diseases that can be used as valuable markers in personalised medicine. Keywords: epistasis, epigenetics, VEGF, APOE, LSR

Résumé Les maladies chroniques, comme les maladies cardiovasculaires (MCV), la maladie d'Alzheimer (AD), la dépression et l'ostéoporose, sont les principales causes de mortalité dans le monde. L'identification de facteurs de risque communs à ces maladies pourrait contribuer à un vieillissement «sain» mieux surveillé en utilisant des stratégies personnalisées de prédiction des risques, de prévention précoce et de traitement adéquat, en tenant compte des comorbidités très souvent existantes. Dans cette thèse, 8 publications ont été développées. Dans un premier temps, j'ai résumé, dans un article de revue, les défis actuels et les opportunités de la pharmacogénomique des médicaments contre les maladies cardiovasculaires. J'ai participé à la formation d'un consortium international, le Consortium VEGF et j'ai participé à une étude qui a identifié des interactions épistasiques entre les polymorphismes qui régulent les niveaux de VEGF et la pression artérielle et les indices d'adiposité. J'ai également démontré qu’un marqueur génétique de VEGF, le rs4416670, était significativement associé à un risque accru de dépression. En outre, j'ai signalé deux interactions significatives entre les variantes liées au VEGF affectant la densité minérale osseuse du col fémoral chez les femmes ménopausées. J'ai également étudié deux marqueurs liés au métabolisme des lipides : l'apolipoprotéine E (APOE) et le «lipolysis-stimulated receptor» (LSR). J'ai trouvé que le variant LSR rs916147 peut interagir avec APOE d'une manière qui inverse l'effet protecteur de l'allèle ε2 de l'APOE sur les lipides sanguins, fournissant ainsi de nouvelles connaissances sur les mécanismes de l'hyperlipoprotéinémie de type III. Les interactions épistasiques entre ces deux gènes augmentent également le risque d’AD même en l'absence de l'allèle à risque, APOE ε4. Finalement, j'ai réalisé des études épigénetiques (EWAS) sur l'obésité centrale et les traits lipidiques chez des individus sains. Les résultats suggèrent qu'un CpG pourrait affecter le tour de taille à travers une voie de signalisation de l'insuline. En outre, deux CpGs ont été associées aux niveaux des triglycérides par des gènes liés aux maladies cardiaques génétiques (PRKAG2) et à l'inhibition de la signalisation Wnt / bêta-caténine impliquée dans le développement des MCV et d’AD (KREMEN2). En conclusion, dans cette thèse j’ai utilisé l'étude de l'épistasie et de l'épigénétique pour identifier des interrelations complexes entre VEGF, LSR, APOE et différentes maladies chroniques (MCV, AD, ostéoporose, dépression) proposant ainsi de nouveaux mécanismes et des dénominateurs communs de ces maladies qui devraient être utilisés comme biomarqueurs de médecine personnalisée. Mots clés: épistasie, épigénétique, VEGF, APOE, LSR

249