Department of Social and Preventive Medicine, Division of Kinesiology-PEPS Building, Laval University, Sainte-Foy, Québec, G1K 7P4, Canada. E-mail: Louis.Perusse@kin.msp.ulaval.ca
This report constitutes the seventh update of the human obesity gene map incorporating published results up to the end of October 2000. Evidence from the rodent and human obesity cases caused by single-gene mutations, Mendelian disorders exhibiting obesity as a clinical feature, quantitative trait loci uncovered in human genome-wide scans and in cross-breeding experiments in various animal models, and association and linkage studies with candidate genes and other markers are reviewed. Forty-seven human cases of obesity caused by single-gene mutations in six different genes have been reported in the literature to date. Twenty-four Mendelian disorders exhibiting obesity as one of their clinical manifestations have now been mapped. The number of different quantitative trait loci reported from animal models currently reaches 115. Attempts to relate DNA sequence variation in specific genes to obesity phenotypes continue to grow, with 130 studies reporting positive associations with 48 candidate genes. Finally, 59 loci have been linked to obesity indicators in genomic scans and other linkage study designs. The obesity gene map reveals that putative loci affecting obesity-related phenotypes can be found on all chromosomes except chromosome Y. A total of 54 new loci have been added to the map in the past 12 months and the number of genes, markers, and chromosomal regions that have been associated or linked with human obesity phenotypes is now above 250. Likewise, the number of negative studies, which are only partially reviewed here, is also on the rise.
This is the seventh update in this series on the status of the human obesity gene map. It incorporates the material published by the end of October 2000. Previous reviews have been published (1) (2) (3) (4) (5) (6). The review has expanded considerably and continues to incorporate evidence from several research approaches, with each approach constituting a section heading. As in previous compendia, the present synthesis includes sections dealing with single human gene mutations, genetically unidentified Mendelian disorders, quantitative trait loci (QTLs) from rodent and other animal model studies, association studies in humans with specific genes and mutations, human linkage studies including genome scans whose goals are to identify QTLs of obesity or obesity-related phenotypes, and a refined pictogram of the 2000 human obesity gene map. The references to each entry in the current human obesity gene map are provided for convenience.
The review includes publications that have dealt with a variety of phenotypes pertaining to obesity, including body mass index (BMI), body fat mass, percentage of body fat, fat-free mass, skinfolds, resting metabolic rates, plasma leptin levels, and other components of energy balance. As in previous updates, negative findings are not systematically reviewed but are briefly introduced when such data were available to us.
In this year's review, we are using gene symbols and chromosomal locations given in the Locus Link website (http:www.ncbi.nlm.nih.govLocusLink) available from the National Center for Biotechnology Information, which provides the official nomenclature of genetic loci.
We renew our request for feedback and comments on the part of the investigators who are interested in any aspects of the gene map and who would like to contribute to future editions of this publication. The electronic version of this publication will be available at the following address: http:www.obesity.chair.ulaval.cagenes.html. An expanded and more extensively cross-referenced version of the map can be accessed at http:www.eureka.pbrc.edu.
We have added to single-gene mutation rodent models (Table 1) the attractin (Atrn) gene from the Mahogany mutation that suppresses diet-induced obesity. In humans, ATRN is located at 20p13 and encodes a transmembrane form of attractin (7).
Table 1. Single-gene mutation rodent models of obesity*
This last year has seen the discovery of another probable gene mutation explaining a human obesity syndrome (Table 2). Holder et al. (8) described a severely obese girl with a weight of 47.5 kg at 67 months of age (+9.3 SD) and a height of 1.27 m (+3.2 SD); no other anomalies were reported. She carried a de novo balanced translocation between chromosomes 1 and 6 (karyotype 46,XX,t(1;6)(p22.1;q16.2)), which disrupted the SIM1 gene, a critical transcription factor for the formation of supraoptic and paraventricular hypothalamic nuclei in mice. The latter nuclei are well known to be involved in energy homeostasis. The patient was heterozygous for a silent C → T substitution at nucleotide (nt) 1328 in exon 9, presumably associated with a loss of function of SIM1. Her energy expenditure was normal for her age and weight, implying that the mutation probably affected energy intake.
Table 2. Cases of human obesity caused by single-gene mutations
Among the previously reported monogenic obesity syndromes, the only new published cases have been on carriers of mutations in the melanocortin-4 receptor (MC4R) gene. The German group that reported on six individuals in 1999 (9) expanded their study to three families, identifying a total of 19 carriers of either a 4-base pair (bp) deletion at codon 211 or a nonsense mutation at codon 35 (10). However, obesity status differed between carriers, suggesting variable penetrance. In a population of 209 severely obese (BMI of >40 kg/m2) French subjects, Vaisse et al. (11) reported on 8 carriers of eight different mutations in the MC4R gene not found in 366 normal weight controls. Several of the mutations appeared to have a functional impact. Thus far, of a total of 47 cases of monogenic forms of obesity involving 19 mutations in six different genes, MC4R mutations seem to be the most frequent, with a total of 11 mutations responsible for 34 cases.
Progress is ongoing in identifying genes or narrowing down the loci associated with the various adiposity-related Mendelian disorders described in the Online Mendelian Inheritance in Man database; a total of 24 syndromes with known map locations are given in Table 3.
Table 3. Obesity-related Mendelian disorders with known map location*
Mode of inheritance
Adapted from OMIM (Online Mendelian Inheritance in Man) computerized database.
N/A, not available.
Status as of October 2000.
Albright hereditary osteodystrophy (AHO)
Albright hereditary osteodystrophy 2 (AHO2)
Angelman syndrome with obesity (AGS)
Insulin resistance syndromes (IRS)
Posterior polymorphous corneal dystrophy (PPCD)
Familial partial lipodystrophy Dunnigan (FPLD)
Prader–Willi syndrome (PWS)
Thyroid hormone resistance syndrome (THRS)
Ulnar-mammary syndrome or Schinzel syndrome (UMS)
Alstrom syndrome (ALMS1)
Bardet–Biedl syndrome 1 (BBS1)
Bardet–Biedl syndrome 2 (BBS2)
Bardet–Biedl syndrome 3 (BBS3)
Bardet–Biedl syndrome 4 (BBS4)
Bardet–Biedl syndrome 5 (BBS5)
Bardet–Biedl syndrome 6 (BBS6)
Berardinelli–Seip congenital lipodystrophy (BSCL)
Cohen syndrome (COH1)
Carbohydrate-deficient glycoprotein type 1a (CDGS1A)
In this year's update, because of the importance of growth hormone on body fat distribution, in particular its association with central adiposity, we have reviewed the hereditary growth hormone and pituitary hormone deficiency syndromes, for which several causal genes have been well-characterized in the last decade. We have only retained those syndromes for which overall or central obesity is clearly specified in the reports.
In the autosomal recessive category, isolated growth hormone deficiencies with mutations in the growth hormone-releasing hormone receptor gene have been described. Wajnrajch et al. (12) reported on two children with predominantly truncal obesity carrying a G → T transversion at nt 265, resulting in a Glu72Stop mutation and a truncated receptor. Salvatori et al. (13) described a G → A transition of nt 1 of the 5′ splice site at the beginning of IVS1 in 30 affected subjects of Brazilian origin with increased abdominal fat accumulation and severe growth retardation.
In the combined pituitary hormone deficiency syndromes, multiple endocrine axes can be affected with variable clinical expression. In particular, hereditary growth hormone deficiency syndrome and central hypothyroidism can lead to truncal obesity. The “Prophet of Pit-1” or PROP1 gene, controlling the ontogenesis of pituitary neuroendocrine cells, is involved in several reported cases of combined pituitary hormone deficiency syndromes. In 1999, Rosenbloom et al. (14) reported on eight subjects with a 2-bp 296delGA deletion, three of whom were over the 90th percentile for BMI with respect to height. Mendonca et al. (15) described two subjects carrying a 2-bp 301delAG deletion, both of whom had marked decreases in height but increases in weight relative to height.
Among the previously reported syndromes, in the autosomal dominant category, the major advance in the last year was the discovery of a gene explaining Dunnigan-type familial partial lipodystrophy, a syndrome in which peripheral subcutaneous fat is absent. Cao and Hegele (16) described five Canadian patients each carrying a novel G → A change at codon 482 in exon 8 (missense mutation R482Q) in the lamin A/C gene (LMNA), which undergoes alternative splicing to produce the nuclear lamin proteins lamin A and lamin C. No unaffected members of the five families carried the mutation.
Regarding other syndromes, several new mutations in the GNAS1 gene were reported by Aldred and Trembath (17), as well as a review of published mutations in patients with Albright hereditary osteodystrophy, a disorder in which obesity is one of the defining features. Prader–Willi syndrome (PWS) remains a subject of intense research, with a French group demonstrating the possible involvement of a novel imprinted gene in the PWS locus, the NDN gene, which is the human homologue of the mouse brain-specific nectin protein gene (18). Another group (19) reported a family with PWS in which affected members carried only submicroscopic deletions in the SNRPN gene and not in nearby loci, therefore confirming previous reports (20) (21) that the PWS imprinting center is located close to that gene.
In the autosomal recessive category, progress was made in Bardet–Biedl syndrome (BBS) patients. Katsanis et al. (22) narrowed down the type 1 (BBS1) locus, through linkage and haplotype analysis in 91 pedigrees of North American and European origin, to a 1.8-megabase region between markers D11S1883 and D11S4944 on chromosome 11. Young et al. (23) further reduced the BBS1 interval to a 1-cM region between markers D11S1883 and D11S4940, surrounding the phosphorylase glycogen muscle locus, in a study of families from Newfoundland, Canada. This same Canadian group also narrowed down the BBS type 3 locus through haplotype analysis to a critical 6-cM interval on chromosome 3 between D3S1595 and D3S1753 (24). Finally, a new BBS, type 6 (BBS6), mapping to 20p12, was described in the past year by two groups (25) (26). Both groups found the gene involved, known as the McKusick–Kaufman syndrome gene. It encodes for a chaperonin protein that plays a role in protein integrity. Affected individuals were either homozygotes or compound heterozygotes for several mutations in the McKusick–Kaufman syndrome gene.
In carbohydrate-deficient glycoprotein syndrome type 1a, a disorder caused by defective glycosylation of glycoconjugates resulting in severe encephalopathy with hypogonadism and lipodystrophy, several new mutations in the phosphomannomutase (PPM2) gene were reported (27) (28). Similarly, in the past year, three different groups (29) (30) (31) described several novel mutations in the solute carrier family 2 gene (SLC2A2), also known as the GLUT2 gene, in patients with Fanconi–Bickel syndrome, in whom there is sparse subcutaneous fat as well as hepatorenal glycogen accumulation.
In the last category, X-linked Mendelian disorders, two new syndromes have been described. Shashi et al. (32) reported on a large American family in which seven males had mental retardation with characteristic dysmorphic features and obesity. Linkage analysis established linkage to Xq26–27. Ahmad et al. (33) described a Pakistani family in which 10 males showed mental retardation, obesity, hypogonadism, and tapering fingers. Maximum linkage was obtained with the marker DXS1106.
Finally, in Simpson–Golabi–Behmel syndrome type 1, in which affected individuals have elevated birth weight and height as well as prenatal and postnatal overgrowth with slight obesity, Veugelers et al. (34) reported on several point mutations and one exon deletion in the glypican-3 (GPC3) gene in seven patients, predicting loss-of-function of the GPC3 protein.
QTLs from Cross-Breeding Experiments
The number of animal QTLs linked to body weight or body fat has increased by 17 since the last review (Table 4). A total of 115 animal QTLs have now been uncovered. Their equivalent syntenic regions in humans, when they can be determined from available maps, are shown in Table 4. When none was provided by the authors, we have proposed acronyms for QTLs to facilitate their inclusion in the map. Overall, the results of one mouse novel cross, two rat novel crosses, and one pig novel cross have been published. Some existing rat and pig crosses have also been further investigated.
Table 4. QTLs reported for animal polygenic models of obesity with their putative syntenic locations in the human genome*
r, replicated; Mob, multigenic obesity; Dob, dietary obese; Afw, abdominal fat weight; Afp, abdominal fat percent; Bl, body length; Bw, body weight; Obq, obesity QTL; Pfat, polygenic fatness; Qlw, QTL late weight gain (6–10 weeks); Qbw, QTL body weight; Qlep, QTL for leptin; Bw6, body weight at 6 weeks; Qfa, QTL LEPRfa; Nidd/gk, type 2 diabetes/Goto–Kakizaki; bw/gk, body weight/Goto–Kakizaki; Niddm, type 2 DM; Hlq, heat loss QTL; Fatq, fat QTL; Batq, brown adipose tissue QTL; Pfatp, predicted fat percentage; Dmo, diabetic mouse; Fob, fat obesity; Wokw, Wistar Ottawa Karlsburg Rt1
W; SHR, salt hypertensive rat; SSC, swine chromosome; AFIFA, avian feed in take at a fixed age interval. p, paternal effect; m, maternal effect; N/A, not available.
Synteny relationships established according to the following references: 84–88.
Also observed in the cross CAST/Ei × C57BL/6J (179).
Also observed in the CAST/Ei × C57BL/6J F2 intercross (193).
Also observed in the cross (C3H/He × A/J) × Mus spretus (65).
A new mouse cross between outbred lines divergently selected for 53 generations for high-fat (fat or F line) or low-fat (lean or L line), and presenting a 5-fold difference in percentage of fat at 14 weeks of age, allowed the detection of four QTLs (Fob1 to Fob4) related to percentage of fat on chromosomes 2, 12, 15, and X (35). Five of the QTLs previously reported for late body weight gain (36) were replicated on chromosomes 4, 5, 7, 9, and 10 in a second F2 intercross of the LG/J and SM/J mouse strains (37). The Wistar Ottawa Karlsburg with RT1u haplotype (W) of the major histocompatibility complex (WOKW) rat strain, which develops the main features of syndrome X such as moderate hypertension, dyslipidemia, hyperinsulinemia, obesity, and impaired glucose tolerance (38), has been crossed to the Dark Agouti/Karlsburg strain. Two new QTLs (WOKW1 and WOKW2) related to body weight and BMI and located, respectively, on chromosome 1 and chromosome 5 (39) have been reported. A QTL for body weight (Dmo1) previously identified from a cross between the diabetic OLETF × BN rat strains (40) was replicated in a backcross to the OLETF strain (41). In addition, new QTLs related to fat weight, body weight, or adipose index on chromosomes 1 (Dmo4), 3 (Dmo5), 6 (Dmo6p), 7 (Dmo7p), and 11 (Dmo9 and Dmo10) were identified.
In a pig cross involving the obese Chinese Meishan breed and lean Dutch White production lines, one QTL (SSCX) for back fat thickness and intramuscular fat content has been detected on chromosome X, near the phosphoglycerate kinase 1 gene (42). In the same cross, paternally or maternally expressed QTLs were also observed for back fat thickness on chromosomes 2 (SSC2) and 7 (SSC7), and two other QTLs (SSC6p, SSC6q) were observed for intramuscular fat on chromosome 6 (43). Finally, in a mix of five different crosses (Meishan × Duroc or Hampshire or Landrace; Minzhu × Hampshire or Landrace) analyzed previously (44), a QTL for back fat thickness at the pituitary specific transcription factor 1 gene, encoding an essential transcriptional regulatory factor of growth hormone, prolactin, and thyrotropin β subunit, was identified on chromosome 13 (45).
We have defined, when not provided by the authors, the putative syntenic relationships with human chromosomes for the QTLs identified in Table 4. To establish the synteny, the position of the markers defining the QTL, according to the Mouse Genome Database from The Jackson Laboratory (Bar Harbor, ME) (46), was compared with the equivalent region in the human genome using the integrated linkage maps of the Mouse/Human homology maps (47). For the rat, the maps described by Yamada et al. (48) and Jacob et al. (49) were used. For the pig and the chicken, maps from the Animal Genome Database in Japan (50) and from the U.S. Livestock Genome Mapping Projects (51) were used.
The evidence for associations between candidate genes and obesity-related phenotypes is summarized in Table 5. Altogether 130 studies have reported significant associations involving a total of 48 candidate genes. Studies published over the past year have shown significant associations of BMI, body weight, and obesity with polymorphisms in TNFRS1B (52), LEPR (53) (54), LMNA (55), peroxisome proliferator activated receptor, gamma (PPARG) (56) (57) (58), UCP1 (59), UCP2 (60) (61) (62), UCP3 (63), NPY5R (64), GRL (65), ADRB2 (66) (67) (68), ADRB3 (69) (70), tumor necrosis factor, alpha (TNFA) (71) (72), NPY (73), LEP (74), INS (75), APOA4 (76), DRD2 (77), GNB3 (78), MC4R (11), LDLR (79), and HTR2C (80). Fat mass and/or percentage of body fat was associated with markers of LEPR (53), CCKAR (81), TNFA (72), UCP2 (60) (61), and APOA4 (76). Polymorphisms in the NPY (82) and GNB3 (83) genes have been reported to be associated with birth weight. Abdominal obesity phenotypes (abdominal visceral and subcutaneous fat, waist girth, waist-to-hip ratio, and abdominal sagittal diameter) showed associations with LEPR (54), LMNA (55), PPARG (56) (57), CART (84), GRL (65), ADRB2 (67), NPY (73), ADRB3 (66), and ADRA2A (66) markers. Polymorphisms in the UCP2 (60), DRD2 (85), and LDLR (79) loci have been reported to be associated with skinfold thickness phenotypes. Plasma leptin levels have been found to be associated with DNA sequence variation in the TNFRS1B (52), LMNA (86), PPARG (57), CCKAR (81), GRL (65), and ADRB2 (87) markers. Finally, a HTR2A gene polymorphism was associated with dietary energy and carbohydrate and alcohol intake in obese subjects (88), and the LMNA gene was associated with familial partial lipodystrophy (89).
Table 5. Evidence for the presence of an association between markers of candidate genes with BMI, body fat, and other obesity-related phenotypes
BW, BMI, % overweight, % fat, FM, 4 skinfolds and their sum
0.001 to 0.05
Obesity (BMI > 30 kg/m2)
BW and FM gain in peritoneal dialysis patients
BMI, RQ, NPRQ, FATOX/LBM in African Americans
0.008 to 0.04
Current BMI, maximal BMI and BW during diet therapy in morbidly obese patients
0.02 to 0.04
BMI in morbidly obese subjects
BMI and WHR in young men without family history of MI
BMI, % fat
0.004 to 0.023
Iliac and triceps skinfold
0.002 to 0.039
Obesity (BMI > 30 kg/m2)
0.002 to 0.003
BMI in hypertensives
Body weight and BMI in young white, Chinese, and black African males
0.001 to 0.05
BMI, waist and hip girths and skinfolds in Nunavut Inuit
BMI in primiparous women
Babies’ birth weight (mothers’ genotype)
FM, % Fat, FFM,
ΔFFM after 20-week endurance training
BMI in healthy lean women
0.004 to 0.03
Dietary energy and carbohydrate and alcohol intake in obese subjects
0.028 to 0.047
BMI in females
FM, % fat, FFM in females
0.002 < p < 0.004
Obesity (BMI > 26) in hypertensives
BMI in hypertensives
BMI in hypertensives
BMI in normotensives
Obesity (BMI ≥ 26)
BMI, triceps and subscapular skinfold, arm fat index
0.001 to 0.021
BMI in type 2 DM subjects
0.0004 to 0.01
BMI > 28 kg/m2
0.009 to 0.02
Gene–gene and gene–environment interactions on obesity-related phenotypes have also been reported. In the Québec Family Study cohort, significant interactions were reported between the ADRA2A and ADRB3 polymorphisms on abdominal total and subcutaneous fat (66). In a group of obese German women, those who carried a rare allele in both the ADRB3 and IRS1 loci (n = 6) lost less weight during a weight loss program than those who were homozygotes for the common allele at both loci (90). In a cohort of Finnish type 2 diabetics and nondiabetic controls, subjects who carried a rare allele at both the ADRB3 and UCP1 loci (n = 11) gained more weight during a 10-year follow-up period than those who were homozygotes for the wild-type allele (91). In a cohort of morbidly obese patients, Otabe et al. (63) reported a significant inverse association between physical activity level and BMI in the homozygotes for the common allele in the UCP3 locus, whereas no association was found in other genotypes. An association between a GNB3 polymorphism and BMI that was detected in primiparous sedentary women was not observed in their physically active counterparts (78).
In addition to the 40 new studies with positive findings, we found 58 studies showing no associations between obesity-related phenotypes and selected candidate genes. Among the negative studies, the most frequent ones were those performed with markers of ADRB3 (six studies) (62) (92) (93) (94) (95) (96), PPARG (seven studies) (97) (98) (99) (100) (101) (102) (103), LPL (four studies) (94) (104) (105) (106), and UCP1 (four studies) (62) (91) (94) (96). Other markers yielding negative findings were related to ADRB2 (107) (108), LEP (109) (110), LEPR (70) (110), LIPC (111), CART (112) (113), TCF1 (114) (115), TNFA (116) (117) (118), UCP2 (119), UCP3 (120), APOA4 (121), DRD2 (64), DRD4 (122), VDR (123) (124), ESR1 (123), BRS3 (125), MC3R (126), MC4R (127), APOE (128) (129), HNF4A (114), IRS1 (90) (114) (130), HNF6 (114), IGF1R (131), IGF1 (131), INSR (132), and adiponectin (133). It should be noted that the majority of the studies with positive findings also reported nonsignificant associations with other obesity-related phenotypes and/or candidate gene markers. Thus, the actual number of negative findings is considerably higher than suggested by the above list.
A summary of the loci that have been shown to be linked to obesity-related phenotypes in genome-wide scans or other linkage studies is presented in Table 6. A total of 59 genes or loci have provided significant evidence of linkage with obesity-related phenotypes. Three genome-wide scans for obesity-related phenotypes were reported since last year's review. One of these genome-wide scans was performed in 193 Finnish obese (BMI of ≥30 kg/m2) sibpairs using 374 markers with an average density of 10 cM (134). The strongest evidence of linkage to obesity was obtained on chromosome Xq24 with DX6804, which had a maximum likelihood score of 3.14. Another obesity QTL was detected on chromosome 18q21, where a maximum likelihood score of 2.42 was obtained between obesity and D18S1155 (134). The results of a genome-wide scan for fat-free mass assessed from underwater weighing measurements were also reported in the Quebec Family Study (135). In the latter study, based on 292 markers with an average intermarker distance of 11.9 cM and typed on a maximum of 336 sibpairs, significant evidence of linkage with fat-free mass was found with three markers (IGF1R, D15S652, and D15S657) on chromosome 15q25-q26. Two other chromosomal regions also provided evidence of linkage with fat-free mass (FFM): one on 18q12 with D18S877 (logarithm of odds [lod] = 3.5) and D18S535 (lod = 3.6) and another on 7p15.3 with D7S1808 (lod = 2.7). Another genome scan for loci linked to plasma leptin concentrations was reported in Pima Indians (113) based on a total of 1199 sibpairs. The strongest evidence of linkage with age- and gender-adjusted plasma leptin concentrations was found on chromosome 6p21 (lod = 2.1) near the marker D6S271 at a map distance of 54 cM from the p-terminal (113). Using the Haseman–Elston sibpair linkage method, the best evidence of linkage with plasma leptin was found with the marker D16S265 on chromosome 16q21 near the BBS2 locus (113).
Table 6. Evidence for the presence of linkage with obesity-related phenotypes
See Appendix for the complete name of the gene. The absence of a gene symbol indicates that the linkage study involved at argeted chromosomal region. QTL means human quantitative trait locus identified from a genome scan.
Status as of October 2000.
p = 0.03
BMI, Σ6 skinfolds, fat mass
0.009 < p < 0.02
Q223R, CA (IVS 3), CTTT (IVS 16)
BMI, Σ6 skinfolds, fat mass, fat-free mass
0.005 < p < 0.05
lod = 2.8
p = 0.04
p = 0.02
p = 0.004
p = 0.02
>5000 Relative pairs
Leptin, fat mass
lod = 4.9/2.8
720 Subjects; 230 families
0.008 < p < 0.03
lod = 7.5
lod = 2.4/2.7
p = 0.03
lod = 2.0
p = 0.02
lod = 2.9
p = 0.03
0.0004 < p < 0.006
p = 0.009
p = 0.02
Triceps, subscapular, suprailiac skinfolds
0.01 < p < 0.03
0.002 < p < 0.05
Suprailiac skinfold, relative weight
0.004 < p < 0.05
lod = 2.1
Principal component of height, weight, skinfolds, abdominal and hip circumferences
In an attempt to replicate the linkage previously reported with obesity on chromosome 10p12 in French families (136), Hinney et al. (137) genotyped 11 markers spanning ∼23 cM from 10p13 to 10q11 in a sample of 386 individuals from 93 German families that had at least two obese children. The marker D10S197, which provided the strongest evidence of linkage in the French study (136), showed a lod score of 1.7 in the German families. In the German study (137), evidence of linkage to obesity was observed with markers D10S204 (lod = 2.0), D10S193 (lod = 1.1), TCF8 (lod = 2.4), and D10S1781 (lod = 2.2), which are ∼5 cM more centromeric than D10S197. These results suggest the presence of an obesity QTL on chromosome 10p11.2 near the centromere. To determine whether genetic variation in the DRD2 gene locus could be responsible for the linkage reported with BMI on 11q24 after a genomic scan in Pima Indians (138), a total of 1187 Pima Indians were genotyped for two polymorphisms within the DRD2 gene (64). The effect of the DRD2 polymorphisms on the 11q24 linkage signal was evaluated by repeating the chromosome 11 linkage analyses and comparing the lod scores obtained from a model that included the DRD2 genotypes with the lod scores obtained from a model that did not include these genotypes. The results of these linkage analyses were the same regardless of whether or not the DRD2 genotypes were included (64), suggesting that the DRD2 polymorphisms are not responsible for the linkage of BMI to chromosome 11q24 in Pima Indians. Finally, a linkage exclusion analysis of the chromosomal region 11q13 containing the UCP2 and UCP3 genes, performed in 458 individuals from 10 randomly ascertained Mexican American families, revealed that genetic variation around the UCP2 and UCP3 genes was unlikely to exert an influence on BMI, fat mass, waist circumference, and plasma leptin levels (139).
Figure 1 depicts the human obesity gene map and incorporates the loci from single-gene mutation rodent models of obesity, human obesity cases due to single-gene mutations, QTLs from crossbreeding experiments and genome-wide scans, all relevant Mendelian disorders that have been mapped to a chromosomal region, and genes or markers that have been shown to be associated or linked with an obesity phenotype. The map reveals that putative loci affecting obesity-related phenotypes are found on all but chromosome Y of the human chromosomes.
The number of genes and other markers associated or linked with human obesity phenotypes continues to expand. The progress made over the last decade is exemplified by the numbers collated in Table 7. It is obvious that the human obesity gene map has become significantly more detailed and complex since the first version developed in 1994. Of course, some of these loci will turn out to be more important than others, and many will eventually be proven to be false positive. The main task, including the identification or the positional cloning of the QTL genes, remains to identify the combination of genes and mutations that are contributing most to human obesity and to define under which environmental circumstances. It is also likely that the topic of gene–gene interactions will receive more attention in the coming years.
Table 7. Evolution in the Status of the Human Obesity Gene Map
The research of the authors on the genetics of obesity is funded by the Medical Research Council of Canada (Grants MT-13960 and GR-15187). C.B. is supported by the George A. Bray Chair in Nutrition. We thank Diane Drolet for her dedicated contribution to the compendium and the development of the manuscript. The list of genes and markers currently in the map as well as the pictorial representation of the map is also available on the website of the Donald B. Brown Research Chair on Obesity at the following address: http:www.obesity.chair.ulaval.cagenes.html and on the website of the Pennington Biomedical Research Center Human Genomics Laboratory at: http:www.eureka.pbrc.edu.