Obesity is marked by an excess fat mass but is also characterized by high phenotype heterogeneity, linked most notably to differences in the stages of weight evolution. There are different stages involved in the development of the disease. At each stage in the development of obesity (weight gain, weight maintenance and chronicisation, variable response to treatment), the manifestation of complications is probably associated with various molecular mechanisms for which most still need to be elucidated . Once obesity is established, it becomes a systemic pathology and a pathology of signals. One of the phenotypes that characterise it is that when you intervene (medical intervention or surgery), subjects generally resist weight loss.
There are many signals playing a role in the physiopathology of obesity. First afferent signals produced by peripheral organs inform the CNS of the status of energy stores.These signals are integrated into the brain which in turn adapts the metabolism to the information provided. One of the best know signals is of course leptin. Signals also permit communication between organs. Adiponectin is an example: it can participate in exchange between adipose tissue and muscle or between adipose tissue and liver. The genetic approach has been proposed as a means to enter into this complex system, by using either the candidate gene approach or using the genome-wide scan approach.
The knowledge brought by the genetic approach is summarized on this slide. First obesity development can be strongly influenced by a limited number of genes in monogenic situations of obesity. The genetic approach in rare obesity cases or syndromes led to the identification of new genes. Then if you go from the left side to the right side of the slide, multiple genes start to interact with environmental factors. So here we have to deal with genetic and phenotypic heterogeneity. Population study and genetic epidemiology have been the approaches to find the causal genes. However, in polygenic forms of obesity, expression gene screening can help to identify the key drivers of obesity physiopathology. The next slides show examples of knowledge brought by the genetic approach.
Bardet-Biedl syndrome is an autosomal recessive disease with the main characteristics of polydactyly, hypogonadism, obesity, mental retardation and pigmental retinitis. Often skeletal deformities such as small stature, skull deformity and dysplastic hip are present as well. The authors described a 10-year-old boy with a classical Bardet-Biedl syndrome which included a great unilateral tibia vara epiphysarea deformity. The dome osteotomy was performed proximally on the tibia, correcting the 40 degrees varus deformity to 9 degrees. The loss of correction was 8 degrees one year after surgery
Although the function of these proteins in the renal epithelium of mammals remains to be fully evaluated, in C. elegans two BBS proteins, BBS-7 and BBS-8, appear to be involved in regulating rates of IFT movement along the cilia axoneme, as defects in these genes result in abnormal ciliogenesis.
A significant success in identifying cases of monogenic obesity stems directly from the study of genes implicated in rodent monogenic obesity. There are at least 200 cases of human obesity which have been associated with a single gene mutation. In all cases, mutation screening of specific candidate genes was conducted in human individuals meticulously characterized by biochemical or hormonal anomalies matching those described in rodent models. The genetic anomalies affect key factors related to the leptin and the melanocortin pathways (Figure 1). This hypothalamic pathway is activated following the systemic release of the adipokine leptin (LEP) and its subsequent interaction with the leptin receptor (LEPR) located on the surface of neurons of the arcuate nucleus region of the hypothalamus. The downstream signals that regulate satiety and energy homeostasis are then propagated via pro-opiomelanocortin (POMC), cocaine-and-amphetamine-related transcript (CART), and the melanocortin system 27. While (POMC) / CART neurons synthesize the anorectic peptide α-melanocyte stimulating hormone (α-MSH), a separate group of neurons express the orexigenic neuropeptide Y (NPY) and the agouti-related protein (AGRP), which acts as a potent inhibitor of melanocortin 3 (MC3R) and MC4R receptors. The nature of the POMC derived peptides depends on the type of endoproteolytic enzyme present in the specific brain region. In the anterior pituitary the presence of the proconvertase-1 (PC1) enzyme produces ACTH and β- lipotropin peptides, whiles the contemporary presence of PC1 and PC2 in the hypothalamus determines the production of α-, β-, γ-MSH and β-endorphins. Mutations were identified in human genes coding for LEP 53, 64, LEPR 11, POMC 40, and PC1 37, 36.
Deletion of a guanine in codon 133 leaves 133 aminoacid codons and a stop codon. The protein is not secreted.
Pro-opiomelanocortin is a protein which is further processed to yield smaller peptides in a tissue-specific way by PROHORMONE CONVERTASES which cut the protein at dipeptide recognition sites. Expression of PC1 at the anterior pituitary level cuts POMC into the peptides shown in blue, while expression of either PC1 or PC2 cuts these peptides into even shorter fragments in the hypothalamus.
In contrast to mutations in leptin, LEPR, POMC and PC1, there are now more than 90 mutations that have been found in the MC4R gene. MC4R mutation-linked obesity represents approximately 2 to 3% of obesity cases. The mutations include frameshift, inframe deletion, nonsense and missense, and are located throughout the MC4R gene.
The role of MC4R mutations in cases of human obesity relies on two main arguments based on the frequency of these mutations in different populations and their in vitro functional consequences. Firstly, MC4R mutations are more abundant in obese populations. Indeed, functional mutations have also been reported in non-obese subjects but to a significantly lesser frequency. Secondly, investigating the molecular mechanisms by which loss of function mutations in MC4R cause obesity have suggested various functional anomalies: abnormal MC4R membrane expression, defective agonist response, and a disruption in the intracellular transport of this protein.
Several lines of evidence support MC4R as an important obesity gene. Its study started because it was a good biological candidate. Invalidation in mouse models leads to obesity (KO). The demonstration of its importance in humans is based on population-based association studies (no linkage) in different countries, co-segregation of genotype and phenotype in families. In addition, functional studies on receptors exhibit reduced function in vitro. Here, the discovery of MC4R-associated mutation validated the rare variant allele hypothesis in obesity.
Recently the group of Steve O’Rahilly described an ARG236GLY mutation that led to early onset obesity (see below). That was not a syndromic form of obesity. They demonstrated that the mutation c o-segregates with obesity in families. Its frequency is higher in obese children. The mutation also has functional consequences. The disruption of the dibasic cleavage site between (beta-MSH) and beta-endorphin reduces its ability to activate MC4R . This important finding raises the hypothesis of the role of POMC gene as a second oligogene. Arg236-to-gly (R236G) missense mutation after POMC gene screening in 230 children The functional loss of both alleles of the human pro-opiomelanocortin (POMC) gene leads to a very rare syndrome of hypoadrenalism, red hair and early-onset obesity. In order to examine whether more subtle genetic variants in POMC might contribute to early-onset obesity, the coding region of the gene was sequenced in 262 Caucasian subjects with a history of severe obesity from childhood. Two children were found to be heterozygous for a mis-sense mutation, R236G, which disrupts the dibasic cleavage site between beta melanocyte-stimulating hormone (beta-MSH) and beta-endorphin. Beta-TC3 cells transfected with the mutant POMC cDNA produced a mutant beta-MSH/beta-endorphin fusion protein. This fusion protein bound to the human melanocortin-4 receptor (hMC4R) with an affinity similar to its natural ligands, but had a markedly reduced ability to activate the receptor. This variant co-segregated with early-onset obesity over three generations in one family and was absent in 412 normal weight UK Caucasian controls. Combining the results in UK Caucasians with a new case-control study in French subjects and three previously published reports, mutations disrupting this processing site were present in 0.88% of subjects with early-onset obesity and 0.22% of normal-weight controls. These results suggest that the R236G mutation may confer an inherited susceptibility to obesity through the production of an aberrant fusion protein that has the capacity to interfere with central melanocortin signaling.
4 MCR4 represents an intermediate situation between rare monogenic and polygenic obesity. Multiple genes may operate in multiple pathways and interact in multiple ways with the environment.
In gene-environment interaction studies, following best-practice recommendations related to sample size, multiple testing and replication will be a challenge. Dealing with large populations, in which the environment is well controlled, is generally very difficult, but can be performed in the European trials.
Classic mutation and epimutation. A classical mutation is associated with a change in the DNA sequence itself. It is usually irreversible, whereas an epimutation corresponds to a heritable change in gene expression that occurs without a change in DNA sequence, affects chromatin conformation and is reversible. Transcriptionally active chromatin regions or euchromatin show a relaxed chromatin while transcriptionally inactive chromatin regions or heterochromatin are far more compacted, with limited access to transcription factors (TF). Epigenetics is an advanced biological system that selectively utilizes genomic information and is involved in various fundamental phenomena. Specifically, it puts emphasis on the regulation of gene expression, through DNA methylation, chromatin, and post-translational modification of proteins such as histones. Nakao M. Epigenetics: interaction of DNA methylation and chromatin. Gene 2001 Oct 31; 278(1-2):25-31.
There is a wide range of environmental factors that may interact with genotype. So far, only about 20 genes have been studied in interaction with diet change or physical activity.
When it comes to finding candidates, we have proven that this can most definitely be done! For example, the identification of new candidate regions has been achieved by the genome-wide scan approach, where the objective is to examine systematically all chromosomes in obese families using polymorphic markers in order to detect increased allele sharing in obese sib-pairs. This task is performed without preconceptions about the functions of the genes and aims to identify known or unknown genes predisposing to obesity. Then powerful molecular tools enable the newly identified genes to be positioned and eventually cloned. It is sometimes challenging to find the exact causative gene when genomic linked regions encompass thousand of bases. The global picture of chromosomal regions linked to obesity illustrates the complexity of this multifactorial disease. More than 200 regions , located on nearly all the chromosomes, have been linked to different obesity-related phenotypes such as fat mass, the distribution of adipose tissue, the occurrence of a metabolic syndrome, resting energy expenditure, energy and macronutrient intake, weight variation, the levels of circulating leptin and insulin.
Population from Maywood are African Americans. The rest are Caucasians.
The complex picture and the resulting inconsistencies mean that it may be better to consider predisposing alleles in term of risk factors. By looking at the complex picture of obesity-related susceptibility genes, it is important to stress that these genetic studies have mainly produced a large repertoire of predisposing alleles whose importance is variable. These variants are not necessarily sufficient to express the obese phenotype . Association of a gene with a complex trait only proposes a factor of risk rather than a role as a causative gene In this table are examples among gene variants we studied more than ten years ago in the French population, with those most recently tested at the bottom. The PPARg gene has been largely studied through meta-analysis. Probably as expected, these studies produce a list of common variants that may modify the risk of disorder occurring but with only a small degree of certainty which has to be considered in the overall picture of other risk factors. This raises the question of the appropriateness of genetic testing in this complex context.
4 MC4R represents an intermediate situation between rare monogenic obesity and polygenic. These genes may operate in multiple pathways and interact in multiple ways with the environment
Gene-environment interaction: polymorphism in PPARG (Luan et al 2001 Diabetes 50:686) This study determined the effects of the peroxisome proliferator–activated receptor (PPAR)-2 Pro12Ala variant on body composition and metabolism and the magnitude of weight regain in 70 postmenopausal women (BMI 25–40 kg/m2) who completed 6 months of a hypocaloric diet. At baseline, BMI, percent body fat, intra-abdominal and subcutaneous abdominal fat areas, resting metabolic rate, substrate oxidation, and postprandial glucose and insulin responses were not different between genotypes (Pro/Pro = 56, Pro/Ala and Ala/Ala = 14). The intervention similarly decreased body weight by 8 ± 1% in women homozygous for the Pro allele and by 7 ± 1% in women with the Ala allele ( P < 0.0001). Fat oxidation did not change in Pro/Pro women but decreased 19 ± 9% in women with the Ala allele ( P < 0.05). Changes in glucose area were not different between groups; however, women with the Ala allele decreased their insulin area more than women homozygous for the Pro allele ( P < 0.05). Weight regain during follow-up was greater in women with the Ala allele than women homozygous for the Pro allele (5.4 ± 0.9 vs. 2.8 ± 0.4 kg, P < 0.01). PPAR-2 genotype was the best predictor of weight regain ( r = 0.50, P < 0.01), followed by the change in fat oxidation (partial r = 0.35, P < 0.05; cumulative r = 0.58). Thus, the Pro12Ala variant of the PPAR-2 gene may influence susceptibility for obesity. (Nicklas, Diabetes 2001)
Physical activity-genotype interaction is illustrated here by the common glutamine glutamic acid variant of adrenergic receptor beta2 that has been studied in many populations. Here Meirhaeghe and colleagues reported that the glutamine variant carrier is associated with obesity but that this effect is seen only in those who are physicaly active.
Here is an example of a common variant in a guanine nucleotide binding protein. Here women carriers of the T allele have higher BMI. This effect is revealed in women with at least one child and is marked in subjects with no physical activity.
So where are going from here? The mission to identify molecular drivers of obesity at the gene and pathway level remains very challenging, but the tools are continuously developing. Gene cloning strategies are improving and new technologies are emerging. In the context of “omic” strategies, the study of the transcriptomics in relation to obesity may be fruitful.
This is Sysiphus. Each time he managed to move his rock up the mountain, the rock fell down and the work had to be done again. I guess this example from Greek mythology can be applied to the task of mapping obesity, where information and knowledge is continuously increasing!
Genetics of Obesity - Diogenes homepage
Genetics of obesity From genetics to functional genomics Prof. Karine Clément Inserm U872 Nutriomique Université Paris 6/Cordelier Research Centre Endocrinology and Nutrition Dept, Pitié-Salpêtrière Paris [email_address]
Genetics of obesity from genetics to functional genomics Contents of presentation Slides Introduction 3 – 5 Monogenic obesity: case study BBS 6 – 21 Other mono- and oligogenic examples 22 – 41 Polygenic obesity: pertinent genes and risk factors 42 – 57 Gene-gene and gene-environment interactions 58 – 63 The future 64 – 69 Abbreviations used 70 - 71
Obesity: chronic disease with different stages of development weight Intervention years Systemic and signal Pathology Constitution Chronic disease Aggravation Resistance/Regain Complications Genes_ Environment Interaction Epigenetic events U872
Bardet-Biedel Syndrome (BBS) Polydactyly Obesity in childhood (75%) Retinitis pigmentosa … . And other diseases (uro-genotal anomalies, kidney malformation) and cognitive dysfunction Discovered in the late 19th centuries Known as a monogenic disease
BBS: ideal case for gene discovery <ul><li>Frequency of the disease was known </li></ul><ul><ul><li>(1 in 150,000 in Europe, higher in Asia/North Africa) </li></ul></ul><ul><li>Monogenic (1 gene, 1 disease well identified) </li></ul><ul><li>Mode of transmission known (recessive) </li></ul><ul><li>Phenotype easy to detect </li></ul><ul><li>Case, Family, samples accessible </li></ul>Statistics models and tools appropriate
1 to 2 years (<6 months) 400-800 Markers Fine Mapping Many Genes Extensive SNP Analysis Therapeutic Target Family Collection Loci 15-30 Mb Loci ~ 5 Mb Microsatellite genotyping > 500 subjects Parents & Children Current Genome scanning approach Family collections Infrastructure: Automated Sequencers PCR machines Liquid handling robots Running costs: High Infrastructure: Automated SNP system PCR machines Liquid handling robots Thousands of samples Running cost: High
Linkage analysis Linkage analysis in complex disease Tests the transmission of alleles from hetero-zygous parents to offspring against H0=1/2 Tests the distribution alleles IBD among affected sib pairs against H0=1/2 Tests the co-segregation of alleles in (large) non-linkage TDT Sib-pair analysis LOD-score analysis ( homozygosity mapping )
Qualitative traits Meth od of sibling pairs (principals) <ul><li>Father AB and Mother CD If the first child is AC </li></ul><ul><li>The second child could be AC AD BC BD </li></ul><ul><li>Number of identical alleles (IBD) 2 1 0 </li></ul><ul><li>Proportion of identical alleles (xi) 1 1/2 0 </li></ul><ul><li>Probability (pi) 1/4 1/2 1/4 </li></ul><ul><li>Proportion of alleles IBD = PIXI = </li></ul><ul><li>Absence of linkage: = ½. If ≥ 1/2 test for linkage (t) </li></ul>CD AB AC ?
BBS: 3 rd surprising result In some BBS families there is a triallelic mode of transmission (genetic epistasis) Homozygous BBS4 Heterozygous Phenotype Adapted from Mutch & Clement, Plos genet 2006 BBS1
BBS: from syndrome to genes and novel pathophysiological mechanisms <ul><li>Oligogenic and not monogenic (12 genes at least; more to be discovered) </li></ul><ul><li>recessive autosomic but also triallelic transmission (12/65 families with BBS mutation have another BBS mutation) </li></ul><ul><li>More heterogeneous than thought </li></ul><ul><li>Opened a new field of research in human pathology : BBS is a ciliopathy </li></ul><ul><li>More to be discovered </li></ul><ul><li>Role in energy regulation </li></ul><ul><li>Gene-phenotype </li></ul><ul><li>Role/mechanisms in common obesity </li></ul>
Strategic « choices » Human Syndromes Gene identification Genome wide scan Molecular/cellular studies Physiology Clinical cases (disease = associated features) Screening of a known gene Gene mutation Hypothesis generating Hypothesis raised Novel disease/Novel syndrome Hypothesis Biochemistry Genetics Comparative genomics
Serum leptin concentration (ng ml -1 ) Ob1 and Ob2 heterozygote sibling normal siblings heterozygote mothers heterozygote fathers 95% confidence intervals of the mean normal children normal adults Inappropriate leptin levels based on corpulence examination
Nature, 387, pp 903-908 June 26, 1997 Homozygous G del codon 133
POMC and derived actions MC1-R MC4-R MC2-R Adrenal gland Hypothalamus Skin Eumelanin pigment Synthesis Feeding inhibition Glucocorticoids POMC ACTH ACTH MSH MSH s s
POMC and derived actions MC1-R MC4-R MC2-R Adrenal Hypothalamus Skin Eumelanin pigment Synthesis Feeding inhibition Glucorticoids POMC MSH <ul><li>Compound heterozygous for exon 3 mutation (G->T nt7013 & del at 7133) </li></ul><ul><li>No ACTH and aMSH synthesis </li></ul>(Krude et al, Nature genet, 1998) ACTH MSH s s
POMC aberrant proteins in humans Adapted from Krude, JCEM, 2003 6922insC
<ul><li>Severe obesity </li></ul><ul><li>Post-prandial Hypoglycemia </li></ul><ul><li>Hypogonadism </li></ul><ul><li>Hypocortisolism </li></ul><ul><li>ProInsulin /Insulin & POMC increased </li></ul><ul><li>Compound heterozygous for 2 mutations (Gly483Arg, exon13, A->C intron4 with deletion exon 5) </li></ul><ul><li>Anomaly of maturation of prohormones (Proinsuline, POMC), </li></ul><ul><li>but also of gut hormones </li></ul><ul><ul><li>(GLPs), leading to intestinal dysfunction </li></ul></ul>Mutation of Proconvertase 1 O’Rahilly et al, NEJM, 1995 &Jackson et al, Nature Genet 1997, Jackson, Nat Genet 2003 Fraction Number
Normal Pathways of Processing and effects of the Putative Defect in Prohormone Convertase 1 in the Study Patient
5* Humans 3* 6* 1* ? Food intake Energy expenditure Leptin and melanocortin mutations Adapted from D Cummings, 2003 Rare syndromes Obesity only Ob Db POMC-/- fat Mc4r (-/-) Obese phenotype Mice
More than 90 mutations in MC4R gene….. 2-3% obesity cases 305 E V F V T L G V I S L L E N I L V I V A I A N K N L S C I L L T I I I T E S G N S V S V L M D A V A F F Y D N V I D S V I C S S L L A S I C S L L S I A D R Y F F T I I T M F F T M L A L M A S L Y V H M F L M A R L H I K R I I G S V T C A A W I C S I I I G V R K V T L F F P A W C V V F V G I L I T L T I A G K M N A G Q R I L N Y A L F N L Y L I L I M C N S I I D P L I Y A H S P M S T T V N I A L Q Y H N M I I I S S Y D A S V I I C L V L P G T A G C P Q N P Y C V C F M S H Y P N S V F L E A S H L R Y S S Q F R S Q E L R K T F K E I I C C L P G L L G C D L S S R Y H L W N R G M H T S L V N S T H R M K G G L S E S C Y G G D S Y D S T F Q A D I L H S I Y F 5 249 10 15 20 25 151 131 120 127 70 63 42 80 90 97 105 137 145 163 174 168 195 179 211 253 242 260 281 298 55 185 267 235 201 216 30 35 40 K V 100 290 308 312 320 H S H L C F* M C S M L I D T P L T D I P W Q -- V T P S E I S S* Y S P H T W S T Q A V* S L K T S R K V L L C R N S S N R NH2 COOH Extracellular Intracellular
Early weight gain in children with MC4R mutations (French children) MC4R homozygote ( AG 346-347) MC4R homozygote (I166V) MC4R heterozygote children (13) Obese children with wild type MC4R (40) Clement Nature 1998 ; Dubern, et al J Pediatr 2001 and 2007 LEPR mutation 97P 50P LepR Mutation LepR Mutation U872
Functional consequences of MC4R mutations AGRP AMPc ? X MSH Food Intake Energy homeostasis Prot Gs + - Data from the French population Membrane expression Receptor activity Genotype-phenotype Relationships AC N C Intracellular retention => 56% of MC4R mutations <ul><li>Deficit of MSH response </li></ul><ul><li>80% of MC4R mutation </li></ul><ul><li>Decreased basal activity </li></ul><ul><li>=> 76% of MC4R mutations </li></ul>Intracellular retention associated with early onset obesity Lubrano-Berthelier et al, HMG 2003 Srinivasan et Lubrano-Berthelier et al JCI 2004 Lubrano-Berthelier et al, JCEM 2006
Evidence that MC4R is an “obesity gene” <ul><li>Biological candidate </li></ul><ul><li>Invalidation in mouse models leads to obesity (KO) </li></ul><ul><li>Population-based association studies </li></ul><ul><li>“ Co-segregation” of genotype and phenotype in families </li></ul><ul><li>Loss of function of variant receptors </li></ul><ul><li>European populations screened </li></ul><ul><li>(>5000 obese screened, >8000 controls) </li></ul><ul><li>UK ( Farooqi &Yeo NEJM 03, Hum Mol Genet 03, JCI 00, Nature Genet 98 ) </li></ul><ul><li>Finland ( Valli-Jaakola , JCEM 04) </li></ul><ul><li>France ( Lubrano, Dubern, Clément Vaisse , Diabetes 04, JCEM 04, Hum Mol Genet 03, J Ped 00, JCI 00, Nature genet 98 , JCEM 06) </li></ul><ul><li>Germany ( Hinney & Hebebrand , Am J Hum Genet 04, JCEM 03, Mol Psy02, Am J Hum Gent 99, Biebermann H, JCEM 06) </li></ul><ul><li>Italy ( Miraglia Del Giudice , JIO 02, Buono 05) </li></ul><ul><li>Spain ( Marti IJO 03) </li></ul><ul><li>Switzerland ( Branson , NEJM 03) </li></ul><ul><li>Denmark (Larsen , JCEM 2005) </li></ul><ul><li>Invalidation in mouse models leads to obesity (KO) </li></ul><ul><li>Population based association studies (no linkage) </li></ul><ul><li>Co segregation of genotype and phenotype in families </li></ul><ul><li>Loss of function of variant receptors </li></ul>Review in Govaerts Peptide 2005
MC4R mutations illustrate the issues raised by predictive medicine in obesity <ul><li>High risk of developing obesity in carriers notably in infancy </li></ul><ul><li>We cannot know when and how </li></ul><ul><ul><li>variable expression, interaction with environment and/or genes, role of Val103Ile as a modulator of the phenotype (Dempfle 2004, Heid 2005) </li></ul></ul><ul><ul><li>Phenotype variation with time </li></ul></ul><ul><li>Or if obesity will develop </li></ul><ul><ul><li>incomplete penetrance </li></ul></ul><ul><li>Functional consequences are heterogeneous </li></ul><ul><ul><li>effect of MC4R powerful agonists ? </li></ul></ul><ul><li>Physicians have to agree about methods of prevention of obesity in the predisposed families ( family counseling ?) </li></ul>
Other oligogenic situations in obesity <ul><li>ARG236GLY mutation in the POMC gene leads to EARLY-ONSET OBESITY in children </li></ul><ul><li>Good biological candidate </li></ul><ul><li>Co-segregates with obesity in families </li></ul><ul><li>Frequency increased in obese (0.88) vs controls (0.22) </li></ul><ul><li>Disruption of the dibasic cleavage site between (beta-MSH) and beta-endorphin reduces its ability to activate MC4R </li></ul><ul><li>Replication is needed </li></ul>Is POMC the second oligogene ? Challis Hum Mol Genet 2002
Treatment of leptin-deficient children Food intake Weight S Farooqi and S O’Rahilly’s group
Leptin treatment in adults homozygous Cys105Thr (Licino, PNAS, 2004) C B A Y 40 35 27 BMI 55 47 51 before WL 76 47.5 60 kg after 18 months treatment rmetHuLeptin (0.01-0.04 mg/kg) daily evening
Methodological caveats A challenge for gene-environment interaction studies <ul><li>Power (increased sample size) </li></ul><ul><ul><li>Major improvement in the last years (data pooling) </li></ul></ul><ul><li>Multiple testing (statistical result corrected) </li></ul><ul><li>Replication (test in independent groups) </li></ul><ul><li>Biological validation </li></ul><ul><ul><li>Functional assessment of putative disease-causing variants </li></ul></ul><ul><ul><li>Evidence for pathophysiological role of the implicated gene </li></ul></ul>Tabor Nat Rev Genet 2002 Cardon Nature Rev Genet 2001 Freely Associating Nature Genet 1999
Variation in the DNA Not reversible --- C --- --- G --- --- A --- --- T --- --- C --- --- G --- Nakao M. Gene.2001 278:25-31. Express differently depending on the combination with the environment Epigenetic Food Nutrient Genes
Complex interactions underlying polygenic obesity Mutch D & Clement K, Plos Genetics 2006 Nutrition Exercise Viruses Social Status Food Abundance Peer pressure Pollution Technological Progress Nutrition Exercise Viruses hormones Social Status Food Abundance Peer pressure Pollution Technological Progress Psychology Psychology Psychology
Identifying Disease Genes Members of a family affected by the Same disease share Identical disease genes Distribution of disease alleles is Different between Cases and Controls Family Based Linkage Studies Population Based Association Studies
Genome wide scan in obesity <ul><li>Europe </li></ul><ul><li>French </li></ul><ul><li>Dutch </li></ul><ul><li>German </li></ul><ul><li>Finn </li></ul><ul><li>British </li></ul><ul><li>… </li></ul><ul><li>North America </li></ul><ul><li>Caucasians US </li></ul><ul><li>Caucasians from Quebec (Quebec family study) </li></ul><ul><li>Mexican & African & Asian Americans </li></ul><ul><li>Amish </li></ul><ul><li>Pima Indians </li></ul><ul><li>Nigerian families </li></ul><ul><li>And others </li></ul>Clement 2002 From the human obesity gene map
Alternative approach - HapMap <ul><li>What is HapMap? </li></ul><ul><ul><li>public resource (www.hapmap.org) </li></ul></ul><ul><ul><li>a catalogue of common (MAF ≥ 0.05) genetic variants that occur in human beings (~1 SNP/1kb) </li></ul></ul><ul><ul><li>genetic data from 4 populations (n = 269) with African, Asian, and European ancestry </li></ul></ul><ul><ul><li>30 trios of Utah residents with European ancestry from the CEPH collection (CEU) </li></ul></ul><ul><li>Aim </li></ul><ul><ul><li>to provide insight into patterns of genetic variation in the human population </li></ul></ul><ul><ul><li>to guide design and analysis of medical genetic studies </li></ul></ul><ul><ul><li>to increase power and efficiency of association studies to medical traits </li></ul></ul><ul><ul><li> Phase II was completed October 2005: > 5.800.000 SNPs </li></ul></ul>
253 QTL / 244 candidate genes (only 22 replicated in 5 independent studies) Polygenic obesity: many loci and over 240 candidate genes Mutch & Clément, PLoS Genetics 2006 ; Rankinen et al 2006 Y 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 X AMERICAN AMISH EUROPEAN AFRICAN PIMA INDIANS ASIAN 2q14.1 Near INSG2 GAD2 ENPP1 SLC6A14 FT0
Positional candidates? <ul><li>Chromosome Xq24. Suviolahti et al , JCI 2003 found association between obesity and an SNP haplotype in the 3'-untranslated region of SLC6A14, an (amino solute carrier family 6 member 14) acid transporter involved in serotonin synthesis and for SNP haplotypes of the SLC6A14 gene (P = 0.0007-0.006) . No recent news about his role or other confirmation </li></ul><ul><li>Chromosome 10 linked locus. Boutin et al , PLOS 2003 : a SNP haplotype, in GAD2, involved in GABA Formation, associated with morbid obesity in French adults. Not replicated in independent population (*4), incl. functional study </li></ul><ul><li>Chromosome 6. Meyre et al . Nature Genet (2005): association between a 3-allele risk haplotype defined by the polymorphisms K121Q, IVS20delT-1, and A-G+1044TGA) and childhood obesity (OR = 1.69), morbid or moderate obesity in adults (OR= 1.50), and type II diabetes (OR = 1.56). ENPP1 is a membrane glycoprotein that inhibits insulin receptor. Not replicated in independent populations </li></ul><ul><li>Chromosome 2. Herbert et al Science 2006 found an association near a SNP upstream the INSIG2 gene associated with common obesity in adults and children. </li></ul><ul><li>Chromosome 16q12 . Scott et al, Frayling et all Science 2007 found an association between the fused toes (FT0) gene and obesity in children and adults. Association confirmed by Dina et al Nat. Genet 2007. Large population discussed but gene role unknown. Replication? </li></ul>
<ul><li>National Heart, Lung and Blood Institute (NHLB1) Framingham Heart Study (FHS), 25y follow-up, heritability 37-54% </li></ul><ul><li>116,204 SNPs in 694 participants, and 86,604 tested for association with BMI </li></ul><ul><li>Keep the top 10 with the highest power estimate </li></ul><ul><li>Only one associates with BMI. Rs756605: CC have 1 unit BMI over GC OR 1.33 [1.20-1.48 ] </li></ul>C Confirmatory analysis in 5 pop
<ul><li>Rs756605 located 10 kb upstream the ATG of INSIG2 (insulin-induced gene) </li></ul><ul><li>INSIG2 Inhibits fatty acid and cholesterol synthesis </li></ul><ul><li>Overexpression of INSIG2 in liver rat decreases TG levels </li></ul><ul><li>Located in a QTL for obesity in mice </li></ul><ul><li>and humans </li></ul><ul><li>But Rs756605 could be in LD with another gene </li></ul>
Risk factors for obesity or related phenotypes hundreds Thousands 1.27-1.35 (Boutin, 2004) Obesity SLC6A14 (risk haplotype) 1.4 (Clément, 1996) High weight gain UCP1 (-3826 A/G)* 3-4 (Clément, 1996) High weight gain In morbid obesity UCP1+ 3-AR 1.30 (Boutin, PLOS, 2003) Morbid obesity GAD2 (risk haplotype) 1.49 (Coudreau, 2004) Obesity dyslipidemia PTP1b (risk haplotype) Odd ratios (risks) Phenotype Gene 1.7 (Clément, 1995) High weight gain 3-AR (Trp64Arg)* 1.53 (Eberlé, 2004) Morbid obesity Diabetes dyslipidemia SREBP (risk haplotype) 1.37 ( Meyre et al, 2005) Diabetes ENPP1 (risk haplotype) 1.5-1.6* (meta-analysis) Diabetes PPAR (Pro12Ala) 1.22-1.67* (3 studies) ( 38,759 participants) obesity FTO gene
Polygenic Individual combination in interaction with environmental factors Rare Monogenics 1 gene 1 disease LEP, LEPR, POMC,PCSK1 SIM,, Increased energy intake ENPP1 Decrease energy expenditure MC4R INS-VNTR MC4R HNF1A inSig SNP FT0 PPARG ADRB3 GYS UCP1 GNB3 APOE AGT KCNJ11
Future Challenge Genes x genes interaction Profile Risk for a given phenotype ? Protective profile U872
24 25 26 27 ≤ 0.39 ≤ 0.51 ≤ 0.66 >0.66 PPARG Pro 12 Ala IMC (Luan et al 2001 Diabetes 50:686) Provided by Pr C Junien Ratio: polyunsaturated fatty acids Saturated fatty acids Genes/ macromolecules Nutrients
Physical activity- Genotype Interaction Role of Adrenergic receptor 2: Gln27Glu Without With physical activity BMI Waist Meirhaeghe Lancet 1999 Provided by C Junien 28 27 26 25 24 100 98 96 94 92 90 88 p < 0.0001 ns p < 0.0001 ns Gln27Gln Glu+/- and +/+
Hormone- physical activity- gene interaction Role of Guanine nucleotide binding protein GNB3 C 825 T T allele C allele Provided by C Junien
Common variant/common disease hypothesis Unsolved questions? <ul><li>Are there common genetic factors specific to obesity? </li></ul><ul><li>What is the influence of common disease-influencing alleles when they are in other genetic backgrounds, in other genetic combinations, influenced by other epigenetic or environmental factors (and how to study them) ? </li></ul><ul><li>If these susceptibility genes are not causative and modify obesity risk in a certain context, what are they doing in the meantime. Are they neutral or deleterious for other diseases? </li></ul><ul><li>Do they have subtle effects in other epigenetic or environmental contexts? </li></ul>Adapted from Becker Medical hypothesis, 2004
<ul><li>Obesity phenotypes (insulin sensitivity) </li></ul><ul><li>Liver diseases </li></ul><ul><li>Asthma </li></ul><ul><li>Psoriasis </li></ul><ul><li>Coeliac disease </li></ul><ul><li>Chronic Bronchitis </li></ul><ul><li>Colitis </li></ul>Example: TNF (G/A –308) functional variant Adapted from Becker Medical hypothesis, 2004 Positive association
Identify key molecular drivers of human obesity C hallenging M ission ? <ul><li>Gene cloning strategies improve </li></ul><ul><ul><li>High density maps (Hapmap) </li></ul></ul><ul><ul><li>SNP mapping (blocks) </li></ul></ul><ul><ul><li>New strategy of analysis in very large populations (SNP mapping) </li></ul></ul><ul><li>« Omic strategies » </li></ul><ul><ul><li>Gen omic </li></ul></ul><ul><ul><li>Transcript omic </li></ul></ul><ul><ul><li>Prote omic </li></ul></ul><ul><ul><li>Metabol omic </li></ul></ul><ul><li>Combined strategies </li></ul>U872
Toward integration of knowledge Large scale expression Animal models New targets? other « Omic » Computational biology Data bases Genetic map U872
Modified from Ritenbaugh C, Kumanyika S, Morabia A, Jeffery R, Antipatis V. IOTF website 1999: http://www.iotf.org SCHOOL WORK, etc Activities Food INTERNATIONAL Globalisation Media A multitude of interacting factors….. LOCAL Agriculture/ market Care Security Transport Food industry Prevention POPULATION % OBESE Infection Work Leisure Family INDIVIDUAL Energy exp Food Energy density NATIONAL/ REGIONAL Education Food Urbanisation Health Social sec’ty Transport Media & Culture Development
Strategy and tool transition Genes Experiments The better the tools become, the clearer the picture…..
Abbreviations used I Insulin receptor IR Interleukin IL Intraflagellar transport IFT Identical by descent IBD Guanine binding proteins (s = stimulating, I = inhibiting) Gs, Gi, Go Glucagon-like peptide GLP Ghrelin receptor GHR Green fluorescent protein GFP Restriction enzymes EcoR1, BspE1 Deoxypyridinoline (bone resorption marker) Dpd Collagen fragment peptide AHDGGR CTX Carboxypeptidase H CPH Carboxypeptidase E CPE corticotropin-like intermediate lobe peptide or ACTH18-39 CLIP Centre d’Etudes du Polymorphisme Humain CEPH Brain-derived neurotrophic factor BDNF Bardet-Biedl syndrome BBS Arculate nucleus ARC Agouti related peptide AGRP Adrenocorticotropic hormone ACTH Acetylcholine AC
Abbreviations used II Number of standard deviations from an age/sex adjusted mean Z-score Tachykynin-related peptide TKRP Transmission disequilibrium test TDT Single strand conformation polymorphism SSCP Single nucleotide polymorphism SNP Drosophila single-minded gene SIM1 Resting metabolic rate RMR Pancreatic Peptide YY 3-36 PYY Paraventricular nucleus PVN Peroxisome proliferative activated receptor, gamma PPARG Pro-opiomelanocortin POMC Prohormone convertase PC (1, etc) Neuropeptide Y NPY Normal NIe Normal N Melanocyte stimulating hormone MSH (α-, etc) Melanocortin4 receptor MC4R Mutant M Lipotropic pituitary hormone LPH Logarithmic odds LOD Leptin receptor LepR