Best Practice & Research Clinical Endocrinology and Metabolism Vol. 15, No. 3, pp. 391±404, 2001
doi:10.1053/beem.2001.0153, available online at http://www.idealibrary.com on
9 Genetics of human obesity Philippe Boutin
Head of Department of Human Genetics CNRS-Institute of Biology of Lille, Pasteur Institute of Lille, France
Professor of Molecular Genetics and Experimental Diabetes, Director Queen Mary and West®eld College, University of London and St Bartholomew's Hospital, London Genome Centre, London, UK
Obesity is a multifactorial condition. Environmental risk factors related to a sedentary lifestyle and unlimited access to food apply constant pressure in subjects with a genetic predisposition to gain weight. The fact that genetic defects can result in human obesity has been unequivocally established over the past 3 years with the identi®cation of the genetic defects responsible for dierent monogenic forms of human obesity: the leptin, leptin receptor, pro-opiomelanocortin, pro-hormone convertase-1 and melanocortin-4 receptor genes. The common forms of obesity are, however, polygenic. The examination of speci®c genes for involvement in the susceptibility to common obesity has not yet yielded convincing results. Approaches involving the candidate genes and the positional cloning of major obesitylinked regions (state-of-the-art future prospects) will be discussed. Key words: genetics; obesity; multifactorial disease; leptin; leptin receptor; POMC; MC4R; genome-wide scans; susceptibility genes for obesity; positional cloning; linkage studies; association studies; linkage disequilibrium.
Obesity is a common disease that has become more prevalent in developed and even in developing countries over recent years.1 About 8±10% of the French population, 17±20% of those in England and Wales and over 25% of North Americans are obese.1±3 In Europe, although obesity is less prevalent in adults than it is in the USA, the prevalence of high weight is increasing among children and teenagers. In France, the latest data show that 16% of children and teenagers are overweight, and the number of obese children has increased ®vefold over the past 10 years.2 Obesity is a risk factor for early mortality and a wide range of metabolic and cardiovascular complications.4 Although the rapid globalization of the Westernised way of life is responsible for the outstanding rise in the number of cases of obesity (about 1 billion subjects now being overweight or overtly obese), obesity is a typical common multifactorial disease in that environmental and genetic factors interact, resulting in a disease state.5 There is strong evidence for a genetic component to human obesity ± the familial clustering (the relative risk among sibs being 3±7)6 and the high concordance of body composition in monozygotic twins7, for example. The role of genetic factors in common obesity is, 1521±690X/01/03039114 $35.00/00
c 2001 Harcourt Publishers Ltd. *
392 P. Boutin and P. Froguel
however, complex, being determined by the interaction of several genes (polygenicity), which may each have a relatively small eect (they are susceptibility genes but not necessary genes) and will work in combination with other(s) and with environmental factors such as nutrition, physical activity and smoking. As complex traits arise via the concerted action of multiple genes (through a network implying genetic heterogeneity and epistatic interactions) with dierent and strong environmental factors, the task of identifying any single susceptibility factor is problematic. MONOGENIC HUMAN OBESITY In contrast to the situation with other complex traits such as diabetes or Alzheimer's disease, human geneticists should be very modest in their claim to have unravelled the genetic basis of obesity in humans. Indeed, despite early evidence that genes are involved in the variation of body fat in man, no obesity gene was identi®ed until the ob gene was discovered in mice.8 Interestingly, even today, most of the syndromic forms of obesity such as Prader±Willi, Cohen, Alstrom and Bardet±Biedl syndromes, have been genetically mapped, but the causative genes have not yet been isolated.9 It is obvious that the extreme rarity of these forms made this search dicult. The alternative strategy used with considerable success was to screen a large number of human subjects for mutations in candidate genes selected on the basis of murine genetic studies. The limitation of this approach is obvious: only previously suspected genes have been investigated. Nevertheless, ®nding mutations in homologous genes causing an overweight phenotype underscored the role of the underlying pathways in energy homeostasis. Even if human monogenic obesity is relatively infrequent, the opportunity to treat certain patients with rare forms of genetic obesity has important implications for the development of similar therapies for the most common forms of obesity. In this respect, there is no doubt that the search for the monogenic paradigms of human obesity was successful as ®ve dierent obesity genes were found in only 2 years: those for leptin and its receptor, pro-opiomelanocortin (POMC), the melanocortin-4 (MC4) receptor and the enzyme pro-hormone convertase-1 (PC1).10±15 Importantly, all the obesity gene-encoded proteins are strongly connected as part of the same loop of regulation of food intake (Figure 1). Leptin (encoded by the human ob-homologous gene), which is secreted by the adipocytes in proportion to their fat content, circulates and binds the long form of the leptin receptor (encoded by the db-homologous gene) in the hypothalamus. POMC gene expression is increased by leptin action, which leads to the production of alphamelanocyte-stimulating hormone (a-MSH), which reduces food intake when it binds to the brain-speci®c MC4 receptor. The fact that no mutation in other genes involved in further pathways potentially regulating energy intake has so far been found in human monogenic obesity may indicate that the leptin pathway is the core of energy balance in humans. Alternatively, it may reveal our ignorance of much of the complex network of proteins regulating body weight. In terms of the latter, one can argue that, despite its initial success, human genetics can still explain only a minority of the mendelian forms of obesity. It is probable that animal studies as well as new genomic approaches will reveal candidate genes that may play a role in monogenic human obesity as well. Monofactorial obesity may be divided into two groups according to the genes involved. The ®rst group comprises very rare recessive forms of obesity that are
Figure 1. Leptin pathway for weight control. All the monogenic obesity gene-encoded proteins (in red) are strongly connected as part of the same loop of regulation of food intake. Ob-R receptor of the leptine; POMC pro- opiomelanocartin; PC1 pro-hormone convertase-1; CART cocaine and amphetamine related transcript; a MSH alpha-melanocyte-stimulating hormone; MCH melanin concentrating hormone; MC4-R melanocortin-4 receptor; Y5-R neuropeptide Y5 receptor.
Genetics of human obesity 393
394 P. Boutin and P. Froguel
associated with pituitary endocrine dysfunction, these being caused by mutations in the leptin, leptin receptor and POMC genes. Two kindreds with a defect in leptin have been reported with loss-of-function mutations.10,16 Homozygous carriers consistently exhibit a phenotype of morbid obesity with an onset in the ®rst weeks of life, increased appetite and hyperphagia, and hypogonadotropic hypogonadism. Treatment with recombinant leptin is fully successful, leading to the recovery of satiety and a dramatic decrease of fat mass without any change in lean mass.17 Only one family with a leptin receptor mutation has been identi®ed.11 In subjects with the homozygous mutation, a truncation of the receptor before the transmembrane domain completely abolishes leptin signalling, leading to a form of massive obesity similar to that seen with leptin de®ciency, along with signi®cant growth retardation and central hypothyroidism. These obese subjects presented, as did their heterozygous parents (who were not obese), with a very high leptin level. Chromatography of the circulating leptin revealed that the hormone was bound to the truncated leptin receptor, leading to an increased plasma leptin half-life, although fat leptin expression was still correlated to fat mass. Therefore, the full knockout of the leptin pathway in humans was not responsible for the compensatory hypersecretion of leptin. Interestingly, the two surviving girls with the mutation were neither diabetic nor hyperlipidaemic. Furthermore, although these teenagers did not show any evidence of puberty at the time of the initial examination, they subsequently became pubertal, indicating that leptin control on puberty is not complete (K. Clement, unpublished data). The key role of the melanocortin system in the control of body weight in humans is evidenced by the discovery of mutations in both the POMC and MC4 receptor genes. Two children with homozygous or compound heterozygous loss-of-function mutations in POMC exhibited a complex phenotype. This re¯ected the lack of pituitary neuropeptides derived from the POMC product, leading to impaired signalling via dierent melanocortin receptors.12 The absence of a-MSH was also responsible for obesity (because of the absence of the melanocortin ligand for the MC4 receptor) as well as for altered pigmentation with red hair (through the MC1 receptor). In addition, the absence of adrenocorticotropin (ACTH) led to adrenal de®ciency via the MC2 receptor. The second group of monogenic forms of non-syndromic obesity is caused by the highly numerous mutations in the MC4 receptor gene.13,14,18 The MC4 receptor gene is the most prevalent obesity gene to date, aecting 1±4% of very obese cases depending on the population. MC4 receptor mutations generally segregate in a family via an autosomal dominant mode of inheritance with variable penetrance. In some consanguinous pedigrees, however, MC4 receptor mutations with a relatively modest loss of function eect appear to be co-dominantly or even recessively associated with obesity. The MC4 receptor-associated human obesity phenotypes are similar to those found in mice lacking MC4 receptors, these showing moderate-to-severe co-dominant obesity (more important in homozygous than in heterozygous knockout mice) with normal neuroendocrine function in terms of adrenal hormones, growth, reproduction and thyroid hormones. In this respect, obesity caused by MC4 receptor mutations is similar to more common forms of obesity with an earlier age of onset and with a trend towards hyperphagia in infancy, a trait that seems to disappear with age. Recent data obtained in mice suggest that impaired MC4 receptor signalling could be involved in hyperinsulinaemia via an impaired negative neuronal control of insulin secretion.19 The MC4 receptor gene might thus be considered to be a `thrifty gene' and a primary target for therapeutic interventions against obesity, regardless of its cause and possibly the metabolic syndrome involved.
Genetics of human obesity 395
COMMON HUMAN OBESITY In contrast to monogenic obesity, the genetic approach to polygenic obesity has so far been less successful. Despite many claims, most of the dozens of investigated genes fail to provide convincing and unambiguous evidence of any involvement in genetic susceptibility to obesity. The reasons for this are numerous, one of the most important being the weakness of our knowledge of the molecular mechanisms of energy balance. In addition, there has been a lack of well-conducted studies on candidate genes in large populations with reliable phenotypes. Two general approaches have been utilized in the search for the genes underlying common polygenic obesity in human. The ®rst focuses on `candidate genes', that is, genes selected as having some plausible role in obesity on the basis of their known or presumed biological role in energy homeostasis. For many years eorts to identify candidate genes for obesity have been concentrated on adipose tissue. In fat, the regulation of thermogenesis by the sympathetic nervous system is mediated by betaadrenergic receptors.20 In humans, beta-3-adrenergic receptors (b3-AR) are modestly expressed in fat and the adipocytes lining the gastrointestinal tract.21 A Trp64Arg mutation located in the ®rst transmembrane domain of the receptor was identi®ed in obese Pima Indians and French and Finnish subjects.22±24 Discordant data were, however, also published25,26, including results on the functional eect of the b3-AR mutation. These studies indicate that the role of this candidate gene in human obesity is, if anything, modest and that it should be examined in relation to other genes in the same pathway. Importantly, in mature brown adipocyte cells, b3-ARs stimulate uncoupling protein1 (UCP1) via a cAMP metabolic pathway. UCPs are inner mitochondrial membrane transporters that dissipate the proton gradient, releasing stored energy in the form of heat.27 An A to G variation in UCP1 was associated with a gain of fat mass in a Quebec family study.28 Additional eects of the G allele of the ÿ3826 variant of UCP1 with the Trp64Arg mutation of the b3-AR gene were shown on weight gain in a morbid obese French population.29 Moreover, polymorphisms in other members of the uncoupling gene family, UCP2 and UCP3, were associated with body mass index in Pima Indians.30 Variations in the b3-AR and UCP genes are probably not sucient on their own to induce obesity. In addition, their function is still being debated, and recent data from UCP2 knockout mice31 have shown no eect on body weight. In contrast, there is now evidence that UCP2 is a potent inhibitor of insulin secretion, these UCP2 knockout mice exhibiting hyperinsulinaemia. All these uncertainties illustrate the complexity of the candidate gene approach, especially when gene function is not understood. Several other candidate gene studies have also been reported. The role of the leptin gene in common polygenic obesity was ®rst suggested by linkage studies.32±34 Polymorphisms within the 50 untranslated region of the human leptin gene were associated with a low leptin level35 and with resistance to a low-calorie diet.36 In this respect, unpublished data from Pakistani families with leptin de®ciency showed that heterozygous relatives have a lower leptin level and a higher fat mass. These data suggest that even a modest reduction of fat-induced leptin secretion may contribute to weight gain (S. O'Rahilly, personal communication, 2000). The peroxisome proliferator-activated receptor-gamma (PPARg) is a nuclear receptor that plays a key role in adipogenesis and may in some respect control the `thrifty gene response' to environmental signals such as nutrients (as fatty acids appear to bind PPARg), leading to ecient energy storage.37 A Pro12Ala variation in the PPARg gene has been shown to be associated with improved insulin sensitivity and
396 P. Boutin and P. Froguel
obesity38, and with a modestly decreased risk of type 2 diabetes (odds ratio 0.85).39 PPARg is a target of thiazoladinediones, agents that are now used in treatment of type 2 diabetes. A new hormone, adipocyte-speci®c resistin, has recently been identi®ed as a protein that is regulated by a thiazoladinedione.40 Moreover, this drug suppresses the expression of resistin through PPARg signalling. Resistin has been shown to induce insulin resistance in adipocyte cells, but knockout experiments and genetic studies in dierent ethnic groups are required to establish the role of resistin and of other fatsecreted proteins in insulin resistance and obesity.41 Unfortunately, the candidate gene approach has to date yielded only putative susceptibility genes with a small or uncertain eect. This lack of power could be explained by the restricted choice of candidates as a result of our ignorance of the pathways disturbed in excess fat storage. However, extensive gene targeting experiments in mice, functional genomics (expression pro®les in dierent tissues of interest in terms of energy balance) and the near completion of the Human Genome Project are providing a new generation of candidate genes for obesity. The choice of an ideal candidate gene may be based on several criteria including: 1. a chromosomal localization near an obesity±linked locus in humans or animal models; 2. the expression pro®le (i.e. in the adipocytes or hypothalamus); 3. expression being regulated by food intake, nutrients or physical activity; 4. gene targeting or overexpression leading to a modi®cation of body weight. The second approach used for identifying genes underlying common polygenic obesity is based on the analysis of genome-wide scans in order to detect chromosomal regions showing linkage with obesity in large collections of nuclear families. This strategy requires no presumptions in terms of the function of genes at the susceptibility loci since it attempts to map genes purely by position. The genotyping of 400 multi-allelic markers (short tandem repeats with a density of 1 marker per 10 cM) enables the identi®cation of regions showing a strong identity by descent in obese family members (i.e. allele sharing in sibships is signi®cantly higher than 50%). Thereafter, susceptibility gene(s) for obesity may be positionally cloned in the intervals of linkage. Five genome-wide scans for obesity genes have been published, these being carried out in Mexican-American families42, French pedigrees43, Pima Indians44,45 and white Americans.46,47 Two papers provided evidence for a candidate region on chromosome 2p21 that could explain a signi®cant part of the variance in leptin level in humans. This linkage was subsequently replicated in a cohort of African-American families.48 A strong candidate gene for obesity, the POMC gene, maps to this region. The POMC gene is expressed in human brain, gut, placenta and pancreas, and is involved in the leptin/melanocortin pathway.49 Moreover, POMC is the precursor of other peptides, including ACTH and MSH, which are involved in energy homeostasis.50 POMC knockout mice51 are obese and have defective adrenal development and altered pigmentation. These features are similar to those of the phenotype of patients with mutations in the coding region of the POMC gene, which are responsible for rare cases of obesity with an early age of onset and with a recessive mode of inheritance.12 In the common forms of obesity, no association was shown between POMC polymorphisms and obesity in Danish and French cohorts.52,53 A polymorphism in the 50 region of POMC was weakly associated with a variation in leptin level in a Mexican± American population54, but this cannot explain the observed linkage at the 2p23 locus. Since the regulatory regions of the POMC gene are still largely unknown, a functional
Genetics of human obesity 397
sequence modifying POMC expression or a nearby other gene may be the aetiological variant responsible for the observed linkage. Other major loci for obesity and leptin levels (Figure 2) were also identi®ed on chromosome 10p11 and on 5cen±q in French families. The 10p locus may account for 20±30% of the genetic risk for obesity in this population.43 Furthermore, the linkage of this chromosomal region to obesity was recently con®rmed in a cohort of obese young German subjects55 as well as in White Caucasians and in African±Americans.56 In addition to this 10p locus, a genome scan performed in white Americans showed evidence for linkage on chromosome 20q13 and on 10q.46 In Pima Indians, the most interesting region of linkage lies on chromosome 11q.44 Comuzzie recently described a new locus at 3q27 that was linked to various quantitative traits characterizing the metabolic/insulin resistance syndrome.47 Interestingly, this 3q27 locus was previously identi®ed as a locus for type 2 diabetes in the French population.57 Several genes map to this region, including the APM1 gene encoding the dierentiated adipocyte-secreted protein ACRP30/adiponectin, which is abundantly present in plasma. The puri®ed C-terminal domain of adiponectin has been reported to protect mice submitted to a high-fat diet from obesity and to rescue obese or lipoatropic mice models from severe insulin resistance. The mechanism appears to involve decreasing the level of plasma free fatty acids and enhancing lipid oxidization in muscle.58 Moreover, the plasma level of adiponectin has been shown to be decreased in obese diabetic subjects59, which makes ACRP30 an attractive candidate gene for fatinduced metabolic syndrome and type 2 diabetes. Further studies will address the role of variations in the ACRP30 gene in obesity and in obesity-associated type 2 diabetes. Although concerns have been raised about the heterogeneity of the obesity phenotype and the reliability of genetic data in multifactorial diseases in general (e.g. the lack of replication), genome scan results in common obesity have been surprisingly reproducible despite dierences in ethnicity and environmental factors. In this context, linkages between obesity and loci on chromosome 2 and 10 have largely been con®rmed, the same, albeit to a lesser extent, applying to the region on chromosome 5. These data imply that, among complex traits, an excess of fat is one of the most inheritable and that a few major loci may contribute to the genetic risk for obesity in human. A logical conclusion would be that obesity may be an oligogenic disease that could be modulated by various polygenic (modi®er genes) and by environmental in¯uences. To identify the true aetiological gene variants associated with the enhanced risk of `typical' obesity, chromosomal regions of linkage should ®rst be re®ned using of a dense map of bi-allelic single-nucleotide polymorphisms (SNPs). Indeed, the state of the art in the positional cloning of complex disorder susceptibility genes implies the systematic use of SNP markers for linkage disequilibrium mapping.60±62 The strength of linkage disequilibrium is quite variable within the genome, ranging from 10 kb to 300 kb or more. It was postulated that working in so-called `isolated' populations would signi®cantly increase the distances of linkage disequilibrium, but recent evidence shows that even working in Finnish or Icelandic populations is not a panacea.63 The identi®cation of the NIDDM1 gene (calpain 10) on chromosome 2q con®rmed that linkage disequilibrium mapping can be a successful strategy to unravel polygenic disease64, but this work also showed the complexity of this approach. In the case of NIDDM1, an intronic polymorphism (UCSNP-43) was associated with type 2 diabetes in Mexican±Americans, three non-coding polymorphisms, including UCSNP-43, subsequently being identi®ed as de®ning an at-risk haplotype. In other ethnic
- Pima Indians; BMI (ref. 44)
- Americans; BMI (refs 46, 56)
Figure 2. Chromosomal location of the obesity loci identi®ed in genome-wide scans studies in dierent populations. BMI body mass index.
American Caucasians; BMI, waist, hip, weight, insulin, insulin:glucose (ref. 47)
Positive linkage in adult obesity
Mexican-American; African-Americans; leptin levels (ref. 42)
German young obese (ref. 55)
French Caucasians; fat mass, BMI and leptin (ref. 43)
398 P. Boutin and P. Froguel
Genetics of human obesity 399
DNA collection Public databases Genotyping technologies
Transcriptional analysis (human and animals)
Identification of susceptibility genes for obesity
- Physiopathology of obesity - Therapeutic targets Figure 3. Identi®cation of susceptibility genes for obesity: an integrated approach utilizing bio-informatic, linkage and association studies, functional genomics and molecular epidemiology. LD linkage disequilibrium.
groups, such as French Caucasians, the rarity of this high-risk haplotype makes it dicult to reach a de®nite conclusion about the role of calpain 10 in type 2 diabetes. Moreover, as the function of this protease is still unclear, this study has emphasized the limitation of genetic studies to prove a functional relation from solely statistics-based methods. To be ecient, linkage disequilibrium mapping should be integrated into a complex positional cloning pipeline combining an extensive family resource, genotyping technologies, functional genomics and molecular epidemiology (Figure 3). Clinical and methodological resources are therefore key to success in the positional cloning of major obesity-linked regions. In each major gene locus for obesity, as de®ned by genome-wide scans studies, all the information from the Sanger centre (http:// www.ensembl.org/Data/) and the Santa Cruz centre (http://genome.ucsc.edu/ goldenPath/hgTracks.html.) will be employed. Researchers will need to take advantage of the vast amount of genomic information, coupled with technical advances in the throughput and accuracy of genotypic analyses. This may allow an identi®cation of transcripts in the linkage disequilibrium regions and the de®nition of their genomic structure and regulatory sequences. A dense map of SNPs will be constructed by in silico SNP detection within existing data-base sequences (or http://ariel.ucs.unimelb.edu.au/cotton/mdi.htm) and by mutation screening around genes that are regarded as strong biological candidates. Linkage disequilibrium mapping strategies implicate an enormous amount of SNP genotyping, requiring a wide variety of techniques to provide rapid, reliable, accurate and inexpensive high-throughput SNPs. Large cohorts of cases and controls, as well as families in which initial linkages have been established, will be necessary for the linkage disequilibrium mapping strategy to detect variants in¯uencing multiple traits and to
400 P. Boutin and P. Froguel
test for gene±gene and gene±environment interactions.65 In addition, at-risk haplotypes will be characterized. Since it is possible that susceptibility to the disease may be caused by a set of polymorphic alleles in linkage disequilibrium, the identi®cation of such haplotypes will be an essential feature of this approach.66,67 Family-based association methods avoid the confounding eects of population structure and also allow the direct observation of haplotypes. Several sibship-based tests could identify linkages and associations between SNPs and/or SNP haplotypes and obesity and/or related traits.68±70 Molecular epidemiological studies in large general populations will provide crucial and complementary information about the broader role of any potential susceptibility genes in obesity. Genetic and functional studies in humans will be used synergistically; tissue pro®ling, for example, may provide the most direct way to improve our overall understanding of the molecular circuitry maintaining energy homeostasis. Expression pro®ling in humans on the one hand, and the genetic analysis of populations on the other, will therefore provide complementary tools to advance our understanding of the complex network of gene±gene and gene±environment interactions underlying the susceptibility to obesity. Following the identi®cation of genetic variations, an exploration of the consequences at tissue level (tissue pro®ling), at organism level and in populations (molecular epidemiology) will clarify the role of these variants in disease pathogenesis and their implications for diagnostic and therapeutic developments. An improved understanding of the genetic and environmental predictors of risk factors provides a rational basis for the strati®cation of the disease risk and the response to treatment, allowing the eective targeting of preventative and therapeutic tools. CONCLUSION The current epidemic of obesity represents a major public health concern given the strong association of adiposity with cardiovascular, metabolic and other morbidities. Preventative and therapeutic approaches have been hampered by a lack of any fundamental understanding of the normal control of human body fat mass and its disturbance in obese states. Geneticists have pioneered the understanding of the genetic basis of obesity through the discovery of the ®rst monogenic defects leading to extreme childhood obesity. The more challenging problem, however, is the identi®cation of the genetic variants that underlie susceptibility to the common forms of human obesity. Based on highquality family material and recent widely con®rmed genome scan results, it is likely that putative aetiological variants in candidate genes will emerge. The successful execution of these studies will require a multidisciplinary approach combining genomics, bio-informatics, expression pro®ling, biochemistry, human physiology and molecular epidemiology. Although the task is considerable, the breadth and depth of expertise now available in the human genetics of complex traits provides a unique opportunity for signi®cant advance. SUMMARY Obesity is a typical, common multifactorial disease in that environmental and genetic factors interact. In certain cases, with severe and childhood-onset obesity, a single gene plays a major role, the environment having a permissive one (role). Rare mutations of
Genetics of human obesity 401
the leptin gene and its receptor, POMC, PC1 and, more frequently, MCR4 receptor (1±4% of very obese cases) have been described. All these obesity gene-encoded proteins are strongly connected as part of the same loop of regulation of food intake. The more common forms of obesity are, however, polygenic. Two general approaches have so far been adopted in the search for the genes underlying common polygenic obesity in human. The ®rst approach focused on genes selected as having some plausible role in obesity on the basis of their known or presumed biological role. This approach yielded only putative susceptibility genes with a small or uncertain eect. The second approach attempts to map genes purely by position and requires no presumption of the function of genes. Genome-wide scans identify chromosomal regions showing a linkage with obesity in large collections of
Practice points . the ®ve genes causing monogenic obesity may be divided into two groups. First, are those related to very rare recessive forms of obesity associated with pituitary endocrine dysfunction, such as the leptin, leptin receptor, PC1 and POMC genes . treatment with recombinant leptin is fully successful, leading to a recovery of satiety and a dramatic decrease in fat mass with no change in lean mass . second, more frequent (1±4% of very obese cases) mutations occur in the MC4 receptor gene: . MC4 receptor-associated human obesity phenotypes show a moderate-to-severe co-dominant obesity with normal neuroendocrine function . obesity caused by MC4 receptor mutations is similar to more common forms of obesity with an earlier age of onset and a trend towards hyperphagia in infancy, a trait that seems to disappear with age . the MC4 receptor gene might be considered to be a `thrifty gene' and a primary target for small molecules employed against obesity in general, regardless of the proximate cause, and possibly against the metabolic syndrome . even if human monogenic obesity is relatively infrequent, the opportunity to treat certain patients with rare forms of genetic obesity has important implications for the development of similar therapies for the most common forms of obesity
Research agenda . our knowledge of the pathways disturbed in excess fat storage needs to be improved . resources from the human genome project will enable the identi®cation of the genes and regulatory sequences enclosed in major chromosomal obesity-linked regions (in humans or animal models) . rapid, reliable, accurate and inexpensive high-throughput SNP detection techniques for linkage disequilibrium mapping and association studies should be developed . there is a need to integrate positional cloning, combining extensive family resources, genotyping technologies, linkage disequilibrium mapping, family-based and population-based association studies, molecular epidemiology and functional genomics
402 P. Boutin and P. Froguel
nuclear families. Genome-wide scans in dierent ethnic populations have localized major obesity loci on chromosomes 2, 5, 10, 11 and 20. Susceptibility gene(s) for obesity may be positionally cloned in the intervals of linkage. The systematic use of SNP markers for linkage disequilibrium mapping strategies, combined with functional genomic studies, will be helpful in providing a better understanding of the molecular determinants of obesity.
REFERENCES 1. World Health Organization. Obesity: Preventing and Managing the Global Epidemic. Report of a WHO Consultation on Obesity. Geneva, 3±5 June, 1997. 2. Maillard G, Charles MA, Tibult N et al. Trends in the prevalence of obesity in children and adolescents in France between 1980 and 1991. International Journal of Obesity 2000; 24: 1608±1617. 3. Prescott-Clarke P & Primatesta P. Health survey for England 1996, London: HMSO, 1998. 4. Lean MEJ, Hans TS & Seidell JC. Impairment of health and quality of life in people with large waist circumference. Lancet 1998; 351: 853±856. 5. Bouchard C. Current understanding of aetiology of obesity: genetic and non genetic factors. American Journal of Clinical Nutrition 1991; 53: 1561±1565. 6. Allison DB, Faith MS, Nathan JS et al. Risch's lambda values for human obesity. International Journal of Obesity 1996; 20: 990±999. 7. Maes HH, Neale MC & Eaves LJ. Genetic and environmental factors in relative body weight and human adiposity. Behavior Genetics 1997; 27: 325±351. * 8. Zhang Y. Positional cloning of the mouse obese gene and its human homologue. Nature 1994; 372: 425±432. 9. Gunay-Aygun M, Cassidy SB & Nicholls RD. Prader±Willi and other syndromes associated with obesity and mental retardation. Behavior Genetics 1997; 27: 307±324. *10. Montague CT, Farooqi IS, Whitehead JP et al. Congenital leptin de®ciency is associated with severe earlyonset obesity in humans. Nature 1997; 387: 903±908. *11. Clement K, Vaisse C, Lahlou N et al. A mutation in the human leptin receptor gene causes obesity and pituitary dysfunction. Nature 1998; 392: 398±401. 12. Krude H, Biebermann H, Luck W et al. Severe early-onset obesity, adrenal insuciency and red hair pigmentation caused by POMC mutations in humans. Nature Genetics 1998; 19: 155±157. *13. Vaisse C, Clement K, Guy GB & Froguel P. A frameshift mutation in human MC4R is associated with a dominant form of obesity. Nature Genetics 1998; 20: 113±114. 14. Yeo GS, Farooqi IS, Aminian S et al. A frameshift mutation in MC4R associated with dominantly inherited human obesity. Nature Genetics 1998; 20: 111±112. 15. Jackson RS, Creemers JW, Ohagi S et al. Obesity and impaired prohormone processing associated with mutations in the human prohormone convertase 1 gene. Nature Genetics 1997; 16: 303±306. 16. Strobel A, Issad T, Camoin L et al. A leptin missense mutation associated with hypogonadism and morbid obesity. Nature Genetics 1998; 18: 213±215. 17. Farooqi IS, Jebb S, Langmak G et al. Eects of recombinant leptin therapy in a child with congenital leptin de®ciency. New England Journal of Medicine 1999; 341: 879±884. 18. Vaisse C, Clement K, Durand E et al. Melanocortin-4 receptor mutations are a frequent and heterogeneous cause of morbid obesity. Journal of Clinical Investigation 2000; 106: 253±262. 19. Cone RD. Haploinsuciency of the melanocortin-4 receptor de®ciency. Journal of Clinical Investigation 2000; 106: 185±187. 20. Nonogaki K. New insights into sympathetic regulation of glucose and fat metabolism. Diabetologia 2000; 43: 533±549. 21. Krief S, Lonnqvist F, Raimbault S et al. Tissue distribution of beta3-adrenergic receptor mRNA in man. Journal of Clinical Investigation 1993; 91: 344±349. 22. Walsotn J, Silver K, Bogardus C et al. Time of onset of non-insulin dependent diabetes mellitus and genetic variation in the beta3-adrenergic receptor gene. New England Journal of Medicine 1995; 333: 343±348. 23. Clement K, Vaisse C, Manning BSJ et al. Genetic variation in the beta 3 adrenergic receptor gene and an increased capacity to gain weight in patients with morbid obesity. New England Journal of Medicine 1995; 333: 352±354.
Genetics of human obesity 403 24. Widen E, Lehto M, Kanninen T et al. Association of a polymorphism in the beta3-adrenergic receptor gene with features of the insulin resistance syndrome in Finns. New England Journal of Medicine 1995; 333: 348±352. 25. Buettner R, Schaer A, Arndt H et al. The Trp64Arg polymorphism of the beta3-adrenergic receptor gene is not associated with obesity or type 2 diabetes mellitus in a large population-based caucasian cohort. Journal of Clinical Endocrinology and Metabolism 1998; 83: 2892±2897. 26. Ghosh S, Langefeld CD, Ally D et al. The W64R variant of the beta3-adrenergic receptor gene is not associated with type 2 diabetes or obesity in a large Finnish sample. Diabetologia 1999; 42: 238±244. 27. Klingenberger M. Mechanism and evolution of the uncoupling protein of brown adipose tissue. Trends in Biochemical Sciences 1990; 15: 108±112. 28. Oppert JM, Vohl MC, Chagnon et al. DNA polymorphism in the uncoupling protein (UCP) gene and human body fat. International Journal of Obesity 1994; 18: 526±531. 29. Clement K, Ruiz J, Cassard-Doulcier AM et al. Additive eects of polymorphisms in the uncoupling protein gene and the beta-3 adrenergic receptor gene on weight gain in morbid obesity. International Journal of Obesity 1996; 20: 1062±1066. 30. Walder K, Norman R, Hanson RL et al. Association between uncoupling protein polymorphisms (UCP2UCP3) and energy metabolism obesity in Pima Indians. Human Molecular Genetics 1998; 7: 1431±1435. 31. Arsenijevic D, Numa H, Pecqueur C et al. Disruption of the uncoupling protein-2 gene in mice reveals a role in immunity reactive oxygen species production. Nature Genetics 2000; 26: 435±439. 32. Clement K, Garner C, Hager J et al. Indication for linkage of the human ob gene region with extreme obesity. Diabetes 1996; 45: 687±690. 33. Reed DR, Ding Y, Xu W et al. Extreme obesity may be linked to markers ¯anking the human ob gene. Diabetes 1996; 45: 691±694. 34. Duggirala R, Stern MP, Mitchell BO et al. Quantitative variation in obesity-related traits and insulin precursors linked to the OB gene region on human chromosome 7. American Journal of Human Genetics 1996; 59: 694±703. 35. Hager J, Clement K, Francke S et al. A polymorphism in the 50 untranslated region of the human ob gene is associated with low leptin levels. International Journal of Obesity 1998; 22: 200±205. 36. Mammes O, Betoulle D, Aubert R et al. Novel polymorphisms in the 50 region of the LEP gene: association with leptin levels and response to low-calorie diet in human obesity. Diabetes 1998; 47: 487±489. 37. Auwerx J. PPARg, the ultimate thrifty gene. Diabetologia 1999; 42: 1033±1049. 38. Deeb S, Fajas L, Nemoto M et al. A Pro12Ala substitution in the human peroxisome proliferatoractivated receptor gamma 2 is associated with decreased receptor activity, improved insulin resistance and lowered body mass index. Nature Genetics 1998; 20: 284±287. 39. Altshuler D, Hirschhorn JN, Klannemark M et al. The common PPARg Pro12Ala polymorphism is associated with decreased risk of type 2 diabetes. Nature Genetics 2000; 26: 76±80. 40. Steppan CM, Bailey ST, Bhat S et al. The hormone resistin links obesity to diabetes. Nature 2001; 409: 307±312. 41. Flier JS. The missing link with obesity. Nature 2001; 409: 292±293. *42. Comuzzie AG, Hixson JE, Almasy L et al. A major quantitative trait locus determining serum leptin levels and fat mass is located on human chromosome 2. Nature Genetics 1997; 15(3): 273±276. *43. Hager J, Dina C, Francke S et al. A genome-wide scan for human obesity genes shows evidence for a major susceptibility locus on chromosome 10. Nature Genetics 1998; 20: 304±338. 44. Norman RA, Thompson DB, Foroud T et al. Genomewide search for genes in¯uencing percent body fat in Pima Indians: suggestive linkage at chromosome 11q21±q22. American Journal of Human Genetics 1997; 60: 166±173. 45. Hanson RL, Ehm MG, Pettitt DJ et al. An autosomal genomic scan for loci linked to type II diabetes mellitus and body-mass index in Pima Indians. American Journal of Human Genetics 1998; 63: 1130±1138. 46. Lee JH, Reed DR, Li WD et al. Genome scan for human obesity and linkage to markers in 20q13. American Journal of Human Genetics 1999; 64: 196±209. *47. Kissebah AH, Sonnenberg GE, Myklebust J et al. Quantitative trait loci on chromosome 3 and 17 in¯uence phenotypes of the metabolic syndrome. Proceedings of the National Academy of Sciences of the USA 2000; 97: 14478±14483. 48. Rotimi CN, Comuzzie AG, Lowe WL et al. The quantitative trait locus on chromosome 2 for serum leptin levels is con®rmed in African-Americans. Diabetes 1999; 48: 643±644. 49. O'Donohue TL & Dorsa DM. The opiomelanotropinergic neuronal and endocrine systems. Peptides 1982; 3: 353±395. 50. Woods SC, Seeley RJ, Porte D et al. Signals that regulate food intake and energy homeostasis. Science 1998; 280: 1378±1383.
404 P. Boutin and P. Froguel 51. Yaswen L, Diehl N, Brennan MB & Hochgeschwender U. Obesity in the mouse model of proopiomelanocortin de®ciency responds to peripheral melanocortin. Nature Medicine 1999; 5: 1066±1070. 52. Echwald SM, Sorensen TI, Andersen T et al. Mutational analysis of the proopiomelanocortin gene in caucasians with early onset obesity. International Journal of Obesity 1999; 23: 293±298. 53. Delplanque J, Barat-Houari M, Dina C et al. Linkage and association studies between the proopiomelanocortin (POMC) gene and obesity in caucasian families. Diabetologia 2000; 43: 1554±1557. 54. Hixson JE, Almasy L, Cole S et al. Normal variation in leptin levels is associated with polymorphisms in the proopiomelanocortin gene, POMC. Journal of Clinical Endocrinology and Metabolism 1999; 84: 3187±3191. 55. Hinney A, Ziegler A & Oener F. Independent con®rmation of a major locus for obesity on chromosome 10. Journal of Clinical Endocrinology and Metabolism 2000; 85: 2962±2965. 56. Price RA, Li WD, Bernstein A et al. A locus aecting obesity in human chromosome region 10p12. Diabetologia 2001; 44: 363±366. 57. Vionnet N, Hani EH, Dupont S et al. Genomewide search for type 2 diabetes-susceptibility genes in French whites: evidence for a novel susceptibility locus for early-onset diabetes on chromosome 3q27± qter and independent replication of a type 2-diabetes locus on chromosome 1q21±q24. American Journal of Human Genetics 2000; 67: 1470±1480. 58. Fruebis J, Tsao T, Javorschi S et al. Proteolytic cleavage product of 30-kDa adipocyte complement-related protein increases fatty acid oxidation in muscle and causes weight loss in mice. Proceedings of the National Academy of Sciences of the USA 2001; 98: 2005±2010. 59. Arita Y, Kihara S, Ouchi N et al. Paradoxical decrease of an adipose-speci®c protein, adiponectin, in obesity. Biochemical and Biophysical Research Communications 1999; 257: 79±83. *60. Terwilliger JD & Weiss KM. Linkage disequilibrium mapping of complex disease: fantasy or reality? Current Opinion in Biotechnology 1998; 9: 578±594. 61. Zollner S & Von Haeseler A. A coalescent approach to study linkage disequilibrium between singlenucleotide polymorphisms. American Journal of Human Genetics 2000; 66: 615±628. 62. Abecasis GR, Noguchi E, Heinzmann A et al. Extent and distribution of linkage disequilibrium in three genomic regions. American Journal of Human Genetics 2001; 68: 191±197. 63. Eaves IA, Merriman TR, Barber RA et al. The genetically isolated populations of Finland and Sardinia may not be a panacea for linkage disequilibrium mapping of common disease genes. Nature Genetics 2000; 25: 320±323. *64. Horikawa Y, Oda N, Cox NJ et al. Genetic variation in the gene encoding calpain-10 is associated with type 2 diabetes mellitus. Nature Genetics 2000; 26: 163±175. 65. Sham PC & Curtis D. Monte Carlo tests for associations between disease and alleles at highly polymorphic loci. Annals of Human Genetics 1995; 59: 97±105. 66. Excoer L & Slatkin M. Maximum±likelihood estimation of molecular haplotype frequencies in a diploid population. Molecular Biology and Evolution 1995; 12: 921±927. 67. Bonnen PE, Story MD, Ashorn CL et al. Haplotypes at ATM identify coding-sequence variation and indicate a region of extensive linkage disequilibrium. American Journal of Human Genetics 2000; 67: 1437±1451. 68. Horvarth L & Laird NM. A discordant±sibship test for disequilibrium and linkage: no need for parental data. American Journal of Human Genetics 1998; 63: 1886±1897. 69. Lunetta KL, Faraone SV, Laird NM et al. Family-based tests of association and linkage that use unaected sibs, covariates, and interactions. American Journal of Human Genetics 2000; 66: 605±614. 70. Abecasis GR, Gordon LR & Cookson WO. A general test of association for quantitative traits in nuclear families. American Journal of Human Genetics 2000; 66: 279±292.