International Journal of Obesity (2000) 24, 93±100
ß 2000 Macmillan Publishers Ltd All rights reserved 0307±0565/00 $15.00
www.nature.com/ijo
Association of sets of alleles of genes encoding
b 3 ± adrenoreceptor, uncoupling protein 1 and
lipoprotein lipase with increased risk of
metabolic complications in obesity
AM Proenza1,2, CM Poissonnet1, M Ozata3, S Ozen4, S Guran4, A Palou2 and AD Strosberg1*
1
Institut Cochin de GeÂneÂtique MoleÂculaire, Laboratoire d'ImmunoPharmacologie MoleÂculaire, CNRS UPR 0415 and Universite de Paris
VII 22, rue MeÂchain, 75014 Paris, France; 2Departmenta Biologia Fonamental i CieÁncies de la Salut, Universitat de les Illes Balears,
Cra. Valldemossa, km 7,5. 07071 Palma de Mallorca, Spain; 3Department of Endocrinology and Metabolism, GulhaÈne School of
Medicine, Etlik-Ankara, Turkey; 4Department of Medical Biology, GulhaÈne School of Medicine, Etlik-Ankara, Turkey
OBJECTIVE: To investigate the relationship between the polymorphisms of the b 3 ± AR (Trp64Arg), UCP1 (A?G) and
LPL (HindIII and PvuII) loci and the metabolic complications associated with obesity in a Turkish population.
SUBJECTS: 271 unrelated individuals of Turkish origin including obese (body mass index, BMI>
>30 kg==m2) and lean
(BMI 25 kg==m2) subjects.
MEASUREMENTS: Anthropometric (weight, height and blood pressure) and metabolic measurements (plasma levels
of glucose, cholesterol and triglycerides), and determination of b 3 ± AR, UCP1 and LPL genotypes by polymerase chain
reaction followed by enzymatic digestion.
RESULTS: The distributions of genotypes for each candidate gene (bb3 ± AR, UCP1 and LPL) were similar between the
obese and the lean subjects. The Arg64 allele of the b 3 ± AR gene was absent from massively obese men. GG carriers of
the A?G variant of the UCP1 gene showed BMI-associated increases of cholesterol levels which were more marked
ˆ0.027) and AG (Pˆ
ˆ0.039) carriers. Obese P‡
‡ carriers of the LPL PvuII variant had signi®cantly higher
than both AA (Pˆ
ˆ0.011), whereas obese P‡
‡P‡
‡ carriers did not have signi®cantly different levels
levels of glucose than non-carriers (Pˆ
ˆ0.087). Moreover, carriers of both alleles (G&P‡
‡) had higher levels of glucose
of triglycerides than non-carriers (Pˆ
ˆ0.048), but did not have signi®cantly different levels of triglycerides than non-carriers (Pˆ
ˆ0.125).
than non-carriers (Pˆ
‡
However, the BMI-associated increase of triglycerides of P‡&G carriers was signi®cantly more marked than that of P‡
ˆ0.0085).
carriers (Pˆ
CONCLUSION: Our data support the idea that alleles of speci®c genes (UCP1, LPL and b 3 ± AR) might play a role in the
development of certain metabolic complications of obesity and might have additive effects when combined with each
other (as in the case of UCP1 and LPL).
International Journal of Obesity (2000) 24, 93±100
Keywords: b 3 adrenoreceptor; uncoupling protein; lipoprotein lipase; DNA polymorphisms; metabolic disorders of
obesity
Introduction
Obesity is associated with several metabolic complications such as hypertension, hypertriglyceridaemia,
hyperglycaemia and insulin resistance, thus constituting a major risk factor for ill health, leading to
cardiovascular diseases and diabetes.1
Metabolic disorders in human beings are complex
multifactorial traits that are in¯uenced by numerous
factors. In addition to the leptin encoded by the OB
gene,2 the products of several other genes have been
shown to be involved in the metabolic network that
controls body weight. The genes encoding the b3
*Correspondence: AD Strosberg, Laboratoire
d'ImmunoPharmacologie MoleÂculaire, Institut Cochin de
GeÂneÂtique MoleÂculaire. 22, rue MeÂchain, 75014 Paris, France.
E-mail: [email protected].
Received 24 February 1999; revised 17 May 1999; accepted
27 July 1999
adrenoreceptor (b3 ± AR) and the uncoupling protein
1 (UCP1) are involved in thermogenesis. Associations
of these genes with obesity have been reported in
various human populations. The Trp64Arg polymorphism identi®ed in the b3 ± AR gene has been
associated with a tendency to lower resting metabolic
rate in Pima Indians3 and with an increased capacity
to gain weight in French patients with morbid obesity.4 In Japanese and Finns, this mutation has been
found to be associated with obesity and hyperinsulinaemia=insulin resistance.5,6 Several other studies
have con®rmed these initial reports while others
have not (reviewed in Strosberg.7 The A?G polymorphism of he UCP1 gene was also studied in
morbidly obese French patients. In this population,
an association with high weight gain during adult life
was observed and an additive effect of both Trp64Arg
and A?G polymorphisms on weight gain was also
reported.8 Nevertheless, none of these genes appear to
play a major role in the development of obesity.
Genetic variants and metabolic disorders in obesity
AM Proenza et al
94
Lipoprotein lipase (LPL) is a key enzyme for the
hydrolysis of triglyceride-rich lipoproteins, and its
activity is positively correlated with plasma cholesterol levels. Various restriction fragment length polymorphisms (RFLP) have been identi®ed at the LPL
locus. The HindIII polymorphism at the LPL gene has
been found to be associated with hypertriglyceridaemia, decreased HDL-cholesterol levels, increased apo
B levels and coronary heart disease.1 Another variant
of the LPL gene, the PvuII polymorphism, has been
associated with variations in plasma triglyceride concentration9 and coronary artery disease severity.10
The products of the LPL, UCP1 and b3 ± AR genes
are metabolically closely related. Sympathetic stimulation of the adipocyte acts in part upon the b3 ± AR,
both in mediating the lipolysis activation and in the
utilisation of intracellular fatty acids in the uncoupled
respiration of the thermogenic mitochondria that
depends on the uncoupling proteins.11 LPL catalyses
the incorporation of fatty acids into the adipocyte,
regulating both availability for thermogenesis and the
ef®ciency of storage as fat.
Numerous data demonstrate a consistent relationship between the presence of obesity and an increased
prevalence of coronary heart disease risk factors such
as hypertension and hyperlipidaemia.12 A recent
survey from the Turkish Heart Study indicates that
the Turkish population exhibits a high incidence of
coronary disease, estimated at 37% of deaths by the
Turkish Ministry of Health.13 These data and the
different genetic background of this population compared with other Caucasian populations made the
Turkish an interesting population for studies on the
relationship between polymorphic forms of candidate
genes and metabolic complications of obesity.
The aim of this study was to investigate the relationship between the polymorphisms of the b3 ± AR,
UCP1 and LPL gene loci and the obese phenotype in a
Turkish population. We tested the hypothesis that
several candidate genes, either individually or combined, may modify metabolic risk variables in obesity,
such as arterial blood pressure, and plasma levels of
glucose, triglycerides and cholesterol. Our data support the idea that alleles of speci®c genes might play a
role in the development of metabolic complications of
obesity and have a possible additive effect when
combined with each other.
Methods
Subjects
The subjects of the present study were recruited at the
Department of Endocrinology, GulhaÈne School of
Medicine, Ankara, Turkey. The whole sample comprised 271 unrelated individuals (186 men and 85
women) of Turkish origin and born in different
regions of the country. From this population we
International Journal of Obesity
constituted two groups: 94 (77 males and 17 females)
lean subjects (body mass index, BMI 25 kg=m2) and
146 (83 males and 63 females) obese subjects
(BMI > 30 kg=m2). The mean age for each group
was 30 1 and 35 1 y, respectively. The control
(lean) subjects were healthy and had no history of
hyperlipidaemia, diabetes or hypertension. In the
obese group, diabetic subjects (n ˆ 40, 27.4% of the
obese subjects) were de®ned according to the oral
glucose tolerance test.14 Diabetic subjects as well as
subjects who were not in fasting conditions when
blood was drawn (n ˆ 21, 9.3% of the population)
were excluded from the genotype ± phenotype relationship study.
Anthropometric and metabolic measurements
Weight and height measurements were recorded. BMI
was calculated as the body weight (kg) divided by the
square of the height (m2). Systolic and diastolic blood
pressure were measured after 10 min rest in the supine
position.
Blood samples were drawn after an overnight fast
and analysed immediately. Blood glucose was measured by an enzymatic colourimetric method using
glucose oxidase15,16 on an RA ± 1000 autoanalyzer.
Total cholesterol was measured by the CHOD-PAP
method using a Menagent Cholesterol-HF kit (Menarini Diagnostics, Florence, Italy).17 Triglycerides were
measured by an enzymatic colourimetric method
using a Menagent Triglycerides kit (Menarini Diagnostics, Florence, Italy).18,19
DNA analysis
Peripheral venous blood (10 ml) was collected in tubes
containing EDTA and kept frozen at ÿ20 C until
processing. Genomic DNA was extracted from white
blood cells by standard methods after digestion with
proteinase K and extraction with phenol=chloroform.20
The detection of the four polymorphisms studied
here was carried out by PCR ampli®cation of speci®c
fragments and subsequent digestion with restriction
endonucleases.
b3 ± AR Trp64Arg polymorphism. PCR ampli®cation
of the region containing the Trp64Arg b3 ± AR polymorphism was carried out. Primers were: forward
primer, 50 CCA GTG GGC TGC CAG GGG 30 ; and
reverse primer, 50 GCC AGT GGC GCC CAA CGG
30 , as previously described.3 The 248 bp ampli®ed
product was digested with BstNI (New England Biolabs, Beverly, MA, USA). The presence of two
restriction sites (Trp64 allele) resulted in fragments
of 97, 61 and 64pb, and the loss of one restriction site
(Arg64 allele) resulted in fragments of 158 and 64pb.
UCP1 A?G polymorphism. For the PCR ampli®cation of the region containing the A?G UCP1 variant,
primers were: forward primer, 5 CTT GGG TAG
Genetic variants and metabolic disorders in obesity
AM Proenza et al
TGA CAA AGT AT 3; and reverse primer, 5 CCA
AAG GGT CAG ATT TCT AC 3.21 The 470 bp
ampli®ed product was digested with BclI (New England Biolabs, Beverly, MA, USA). The presence of
the restriction site (G allele) resulted in fragments of
250 and 220 bp.
LPL HindIII polymorphism. For the detection of the
HindIII polymorphism of the LPL gene, a 1.3 kb
fragment of genomic DNA was ampli®ed. Primers
were 50 TTA GGC CTG AAG TTT CCA C 30 and 50
CTC CCT AGA ACA GAA GAT C 30 .22 After
digestion with HindIII (GibcoBRL, MA, USA), the
H‡ allele resulted in two fragments of 700 and
600 bp, whereas Hÿ allele retained the size of 1.3 kb.
LPL PvuII polymorphism. For ampli®cation of the
sequence around the PvuII site the primers were 50
TAG AGG TTG AGG CAC CTG TGC 30 and 50 GTG
GGT GAA TCA CCT GAG GTC 30 .22 The 858 bp
ampli®ed fragment was digested with PvuII (GibcoBRL, MA, USA). The presence of the PvuII site
(P‡ allele) yielded fragments of 592 and 266 bp.
Statistical analysis
Statistical analysis was performed using the chisquare test for comparisons of genotype frequencies.
A test of linear trend was done with the b3-AR
Trp64Arg polymorphism. Allele frequencies for
each polymorphic site were estimated by the genecounting method. The distribution of single diallelic
RFLPs was tested for Hardy ± Weinberg equilibrium
with a chi-square test. For comparison between groups
Student's t-test, one-way ANOVA and multiple linear
regression analysis were performed. All statistics were
performed using the BMDP and SPSS-X packages on
a VAX8820 computer.
Results
Table 2 Distribution of genotypes of b3 ± AR Trp64Arg mutation,
UCP ± 1 A?G variant and LPL HindIII and PvuII polymorphisms in
lean and obsese subjects
Lean group
b3 ± AR Trp64Arg Trp64Trp64
Trp64Arg64
Arg64Arg64
UCP A?G
AA
AG
GG
LPL HindIII
H‡H‡
H‡Hÿ
HÿHÿ
LPL PvuII
P‡P‡
P‡Pÿ
PÿPÿ
95
Obese group
n
%
n
%
w2
Pa
74
18
1
49
33
12
55
31
4
31
31
17
79.6
19.4
1.1
52.1
35.1
12.8
61.1
34.4
4.4
39.2
39.2
21.5
115
13
2
65
59
12
75
49
9
37
47
22
88.5
10.0
1.5
47.8
43.4
8.8
56.4
36.8
6.8
34.9
44.3
20.8
4.0
0.135
1.99 0.370
0.79 0.674
0.52 0.770
a 2
w -test obese vs lean subjects.
surprisingly, subjects with BMI values over 30 have
signi®cantly higher glucose, cholesterol and triglyceride concentrations than lean subjects.
Allele and genotype frequencies
Allele frequencies for all polymorphisms were in
Hardy ± Weinberg equilibrium in both the lean
(Arg64 allele frequency: 0.11; G allele frequency:
0.30; H‡ allele frequency: 0.78; P‡ allele frequency:
0.59) and obese (Arg64 ellele frequency: 0.07; G ellele
frequency: 0.31; H‡ allele frequency: 0.75; P‡ allele
frequency: 0.57) groups. In general, we found that the
distributions of genotypes for each candidate gene
(b3 ± AR, UCP1 and LPL) were similar between the
obese patients and the lean subjects (Table 2). However, it can be seen that the percentage of men
carrying the Arg64 allele (both in heterozygote and
homozygote form) decreases as obesity degree
increases (Figure 1), and reaches the zero value for
morbid obesity (test of linear trend: chi-square ˆ 5.78,
P ˆ 0.016). Interestingly, the percentage of women
carrying the Arg64 allele does not change in the
different obesity groups (test of linear trend: chisquare ˆ 0.80, P ˆ 0.37).
Metabolic parameters in lean and obese subjects
Two groups of individuals were studied Ð lean
(BMI 25) and obese (BMI > 30). Table 1 presents
clinical characteristics for both study groups. Not
Table 1 Anthropometric data and plasma levels of glucose,
total cholesterol and triglycerides in lean and obsese subjects
BMI (kg=m2)
Systolic blood pressure (mmHg)
Diastolic blood pressure (mmHG)
Glucose (mmol=l)
Cholesterol (mmol=l)
Triglycerides (mmol=l)
Lean group
(n ˆ 86)
Obese group
(n ˆ 92)
22.3 0.2
114 1
66.5 0.8
4.38 0.07
4.16 0.10
1.62 0.07
37.8 0.5*
132 2*
83.2 1.5*
5.27 0.11*
5.30 0.13*
2.36 0.11*
Values are mean s.e.m. *P < 0.05 obese vs lean.
Figure 1 Prevalence of the Arg64 allele of the b3 adrenoceptorTrp64Arg polymorphism in both men and women according
to the degree of obesity (n ˆ 162 men and 70 women). Test of
linear trend: *P ˆ 0.02; DP ˆ 0.37.
International Journal of Obesity
Genetic variants and metabolic disorders in obesity
AM Proenza et al
96
Genotype ± phenotype relationship in lean and obese
subjects
We observed (Figure 2) that there was a liner, positive, relationship between plasma levels of cholesterol
and BMI values for each of the three UCP1 genotype
groups Ð AA (P ˆ 0.0001), AG (P ˆ 0.0045) and GG
(P ˆ 0.0000). Moreover, comparisons between lines
indicate differences between GG subjects and both
AA (P ˆ 0.027) and AG (P ˆ 0.039) subjects. Thus,
carriers of the G allele in homozygote form showed
signi®cantly more marked (slop ˆ 5.6) increases of
cholesterol levels associated with the degree of obesity than both non-carriers (slope ˆ 2.0) and carriers
of the mutated allele in heterozygote form
(slope ˆ 1.9). There were no differences between
AA and AG lines. Comparisons between groups (ttest) indicated differences in the levels of cholesterol
between obese AA carriers and GG carriers that did
not reach statistical signi®cance (P ˆ 0.22).
In the obese group, carriers of the LPL PvuII P‡
allele (Figure 3), in both homozygote and heterozygote form, had higher levels of glucose than noncarriers (P ˆ 0.066 and P ˆ 0.011, respectively).
There was a similar trend for triglyceride levels
(Figure 4), although it was not signi®cant (P ˆ 0.087
and P ˆ 0.166, respectively).
Obese subjects carrying both the P‡ and G alleles
(Figure 5 had signi®cantly higher levels of
glucose than non-carriers (P ˆ 0.048). No signi®cant
Figure 3 Plasma levels of glucose in lean and obese subjects
according to the LPL PvuII genotype. *P < 0.05 obese vs lean;
s
P < 0.05 obese P‡Pÿ vs obese PÿPÿ.
Figure 4 Plasma levels of triglycerides in lean and obese subjects according to the LPL PvuII genotype. *P < 0.05 obese vs
lean.
Figure 2 (A) Plasma levels of cholesterol in lean and obese
subjects according to the UCP1 A?G variant genotype. *P < 0.05
obese vs lean. (B) Plasma levels of cholesterol in Turkish subjects according to their BMI values. AA y ˆ 119‡2.0x; AG:
y ˆ 128‡1.9 x; GG: y ˆ 17‡5.6x *P < 0.05 GG line vs AA line;
d
P < 0.05 GG line vs AG line.
International Journal of Obesity
Figure 5 Plasma levels of glucose in lean and obese subjects
according to the UCP1 A?G and LPL PvuII genotypes. *P < 0.05
obese vs lean; P < 0.05 obese G&P‡ carriers vs obese noncarriers.
Genetic variants and metabolic disorders in obesity
AM Proenza et al
Figure 6 (A) Plasma levels of triglycerides in lean and obese
subjects according to the A?G UCP1 and LPL PvuII genotypes.
*P < 0.05 obese vs lean; (B) Plasma levels of triglycerides in
Turkish subjects according to their BMI values. P‡:
y ˆ 99‡2.4x; G&P‡: y ˆ ÿ34‡7.6x. *P < 0.05 G&P‡line vs P‡line.
difference between other genotype groups was found,
thus indicating that the magnitude of the increase of
the levels of glucose in obese subjects was associated
with the presence of both alleles.
Carriers of both alleles (P‡ & G) also tended to
have higher levels of triglycerides (Figure 6) than
non-carriers, although this difference did not reach
statistical signi®cance (P ˆ 0.125). Multiple regression analysis indicated that there was a linear relationship between levels of triglycerides and BMI in
subjects carrying the P‡ allele (slope ˆ 2.4,
P ˆ 0.018) and also in subjects carrying the G & P‡
alleles (slope ˆ 7.6, P ˆ 0.0000), and that both lines
signi®cantly differed (P ˆ 0.0085).
No signi®cant difference in the levels of glucose,
cholesterol and triglycerides between genotype groups
was found in lean subjects.
Discussion
To the best of our knowledge, the present data provide
the ®rst report of obesity candidate gene analysis in
Turkey. This study suggests that alleles of speci®c
genes, both individually and in combination, may
intervene in the worsening of the metabolic complications of obesity.
A problem frequently encountered in association
studies is the dif®culty in selecting appropriate controls. This explains the distortion of the results for a
speci®c genetic marker in many case ± control studies.
In the present paper, both obese and controls were of
Turkish ancestry and came from the same geographic
regions, which reduces biases associated with differences in ethnic or genetic backgrounds. In addition,
the genetic homogeneity of the Turkish population has
previously been reported.23
The genotype frequencies of the b3 ± AR, UCP1 and
LPL gene polymorphisms found in our sample are
similar to those previously reported in other Caucasian populations1,4,22,24,25 and do not differ when lean
and obese groups are compared. These genes taken
separately do not appear to constitute major determinants for the development of obesity.
Interestingly, the Arg64 allele of b3 ± AR gene is
absent in massively obese men. CleÂment et al4 indicated that the Trp64Arg mutation may have a deleterious effect because in their study, among 185 patients
with morbid obesity, 14 were hetrozygous and none
were homozygous. Recent review on Trp64Arg polymorphism including over 10,000 patients suggests a
decreased life expectancy of Arg64 homozygotes,
except in Pima Indians, given the much lower than
expected prevalence in the other populations.7 In the
Turkish population, the number of homozygotes is
also low (n ˆ 3). Moreover, the proportion of Arg64
carriers, both homozygotes and heterozygotes,
decreases as BMI increases in men but not in
women. Our data contribute to the idea of a possible
selective effect of this mutated allele directed to
morbidly obese men. Cardiovascular complications
of obesity are more common in men than women,
and several risk factors, such as hypertension, hyperlipidaemia and glucose intolerance, have been thought
to be caused by the increased release of free fatty
acids by the portal venous system. b3 ± AR in visceral
fat appears to be responsible for lipolysis and fatty
acid release in the portal vein26 and the rate of this
catecholamine-induced mobilisation is higher in men
than in women.27 It could be hypothetically considered that the decrease in the proportion of male
carriers of the Arg64 allele could re¯ect the genderspeci®c differences in metabolic and cardiovascular
disturbances that accompany obesity. However to
support this assumption, further studies are needed
in the present or other human populations.
In our population the ratio between women and
men is higher in the obese group than in the lean
group. In the obese group 44% are women, whereas
there are only 18% in the lean group. However, this
fact does not affect our results since there are no
signi®cant differences in the levels of glucose, cholesterol and triglycerides between men and women in
97
International Journal of Obesity
Genetic variants and metabolic disorders in obesity
AM Proenza et al
98
the obese group when considering the whole population and, when analysing the group excluding diabetics, although women showed 11% higher glucose
levels than men this fact does not alter the differences
between genetic variants found for this parameter.
Actually, to avoid possible false conclusions due to
the presence of diabetes, normoglycaemics have been
analysed separately. Interestingly, when comparing
both subpopulations (diabetics vs normoglycaemics)
the distributions of the Trp64Arg b3 ± AR genotypes
are statistically different between groups (w2 ˆ 7.8,
P ˆ 0.0207), the Arg64 carriers being more represented in the diabetic group. Thus, 8.3% of the nondiabetic subjects are carriers of the Trp64 allele (in
heterozygote form only), whereas the percentage is
21.9% in the diabetic group (15.6% of heterozygotes
and 6.3% of homozygotes). This would give further
support to the proposed involvement of the genetic
variants studied in the metabolic complications of
obesity. However the limited number of diabetics
(n ˆ 32) prevents us from reaching more detailed
conclusions. Apart from this quotation the diabetic
subjects were not considered in the genotype ± phenotype analyses but rather excluded.
Our results indicate that among the polymorphisms
studied the A?G variant of the UCP1 gene and the
PvuII polymorphism of the LPL gene appear to be the
more clearly involved in the metabolic disorders
linked to obesity. The presence of the mutant allele
G, mainly in its homozygote form, is statistically
associated with a higher increase of plasma cholesterol levels, an increase which is signi®cantly higher
than that observed in the non-carriers of this allele.
Cholesterol is among the risk factors for cardiovascular disease which have been associated with obesity
by epidemiological studies.28 Moreover, changes in
body weight have been found to be associated with
changes in blood levels of cholesterol.29,30 Our study
suggests that the intensity of these changes would
increase as the presence of UCP1 G allele does.
The association between LPL PvuII P‡ allele and
the higher increase of triglyceridemia of obesity found
here agrees with that reported by a previous work.9
This relationship could have a physiological importance since increased triglyceride levels in women
have been found to be associated with higher waist-tohip ratio Ð a measure of central adiposity.31 Moreover, central adiposity strongly correlates with cardiovascular disease.32 Thus, the P‡ allele would be
involved in the metabolic complications of obesity
due to its effect on triglyceride and glucose levels.
Our results did not show a signi®cant effect of the
H‡ allele on obesity-related metabolic complications
(data not shown) in spite of the fact that both alleles
(P‡ and H‡) belong to the same gene. Discrepancies
between the effects of two variants (HindIII and
PvuII) which are present in the same gene have
already been reported in other studies. Gerdes et al24
found an association between the H‡ allele and low
HDL-cholesterol levels attributable to the H‡Pÿ
International Journal of Obesity
haplotype, thus suggesting that the HindIII site is in
linkage disequilibrium with a region in or near the
LPL gene, affecting the lipolitic function of LPL. Also
that the mutation=polymorphism could have occurred
on an H‡ chromosome before the mutation, causing
the PvuII site variation. Moreover, Mitchell et al33
indicated that, as both polymorphisms are located in
introns, they are unlikely to directly affect either LPL
protein structure or its regulation. Thus, the most
likely explanation for the association between these
variants and quantitative variation in lipid levels in
that the HindIII site is in linkage disequilibrium with
one or more functional polymorphisms of the LPL
gene that directly mediate HDL cholesterol and triacylglyceride levels, whereas the PvuII site would not
be in linkage disequilibrium with the putative functional mutation. Either way, it must also be considered
that in our study few individuals have the HÿHÿ
genotype due to the higher frequency of the H‡ allele
(0.75) compared with the P‡ allele (0.57), which
explains the dif®culty in reaching statistical signi®cance between genotype groups.
There is a combined effect of the G and P‡ alleles
on the obesity-linked increase of the levels of glucose
and triglycerides. In addition, our results suggest that
when obese subjects are carriers of both G and P‡
alleles he increase of triglyceridaemia is more marked
than that produced by P‡ allele alone (see Figure 6).
Thus, the A?G variant, in enhancing the effect of P‡
allele on triglyceride levels, could play a role in
obesity-linked lipid disturbances which have been
classically restricted to LPL gene polymorphisms.
Genetic predisposition has been hypothesized to be
a key factor in the susceptibility of obese individuals
to metabolic disorders.34 Our results reveal that the
consequences of body weight increase on he metabolic parameters in our population vary according to
the genotype. This fact supports the idea that the
genes studies, UCP1, LPL and b3 ± AR to a lesser
extent, could be involved in a susceptibility towards
obesity-linked metabolic alterations. It can be noticed
that the differences found here between genotype
groups, when existing, are always present in the
obese group only and not in the lean group, suggesting
that these are obesity-conditioned effects and not
direct effects of the genetic variants on the metabolic
parameters. Further studies, including more candidate
genes and other populations, are needed in order to
con®rm these results. As the combination of G and P‡
alleles is here speci®cally linked to glycaemia and
hypertriglyceridaemia when obesity progresses, it is
of interest to analyse the effect of each polymorphism
not only taken independently but in combination with
each other.
In conclusion, our data support the idea that alleles
of speci®c genes (UCP1, LPL and b3 ± AR) might
play a role in the development of metabolic complications of obesity and have possibly additive effects
when combined with each other (as in the case of
UCP1 and LPL).
Genetic variants and metabolic disorders in obesity
AM Proenza et al
Acknowledgements
We are extremely grateful for the helpful discussion
with Dr France PieÂtri-Rouxel and for the excellent
technical assistance of Mrs Annabelle Amiard. We
also gratefully acknowledge Mrs Yolande Aron for
experienced technical advice. Support for our work
comes mostly from the Centre National de la
Recherche Scienti®que, the University of Paris VII,
the French Ministry for Education, Research and
Technology and grant CHRX-CT94-0490 from the
European Union. We are also grateful for help from
the Ligue Nationale contre le Cancer, the Fondation
pour la Recherche MeÂdicale FrancËaise, grant
DGICYT PB94-1178 from the Spanish Government
and, last but not least, the Association pour la
Recherche contre le Cancer.
References
1 Vohl MC, Lamarche B, Moorjani S, Prud'homme D, Nadeau
A, Bouchard C Lupien PJ, Despre JP. The lipoprotein lipase
HindIII polymorphism modulates plasma triglyceride levels in
visceral obesity. Arterioscler Thromb Vasc Biol 1995; 15:
714 ± 720.
2 Zhang Y, Proenca R, Maffei M, Barone M, Leopold
L, Friedman JM. Positional cloning of the mouse obese
gene and its human homologue. Nature 1994; 372: 425 ±
432.
3 Walston J, Silver K, Bogardus C, Knowler WC, Celi FS,
Austin S, Manning BS, Strosberg AD, Stern MP, Raben N,
Sorkin JD, Roth J, Shuldiner A. Time of onset of non-insulindependent diabetes mellitus and genetic variation in the b3adrenergic-receptor gene. New Engl J Med 1995; 333: 343 ±
347.
4 CleÂment K, Vaisse C, Manning BS, Basdevant A, Guy-Grand
B, Ruiz J, Silver KD, Shuldiner AR, Froguel P, Strosberg AD.
Genetic variation in the b3-adrenergic receptor and an
increased capacity to gain weight in patients with morbid
obesity. New Engl J Med 1995; 333: 352 ± 354.
5 Kadowaki R, Yasuda K, Iwamoto K, Otabe S, Shimokawa
K, Silver K, Walston J, Yoshinaga H, Kosaka Y, Yamada
N, Saito Y, Hagura R, Akanuma Y, Shuldiner A, Yazaki
Y, Kadowaki T. A mutation in the b3-adrenergic receptor
gene is associated with obesity and hyperinsulinemia in
Japanese subjects. Biochem Biophys Res Commun 1995;
215: 555 ± 560.
6 WideÂn E, Lehto M, Kanninen T, Walston J, Shuldiner AR,
Groop LC. Association of a polymorphism in the b3-adrenergic-receptor gene with features of the insulin resistance syndrome in Finns. New Engl J Med 1995; 333: 348 ± 351.
7 Strosberg AD. Association of b3-adrenoreceptor polymorphism with obesity and diabetes: Current status. Trends Pharmac
Sci 1997; 18: 449 ± 454.
8 CleÂment K, Ruiz J, Cassard-Doulcier AM, Bouillaud F,
Ricquier D, Basdevant A, Guy-Grand B, Froguel P. Additive
effect of A?G (-3826) variant of the uncoupling protein gene
and the Trp64Arg mutation of the b3-adrenergic receptor gene
on weight gain in morbid obesity. Int J Obes 1996; 20: 1062 ±
1066.
9 Chamberlain JC, Thorn JA, Oka K, Galton DJ, Stocks J. DNA
polymorphisms at the lipoprotein lipase gene: Associations in
normal and hypertriglyceridaemic subjects. Atherosclerosis
1989; 79: 85 ± 91.
10 Wang XL, McCredie RM, Wilcken DEL. Common DNA
polymorphism at the lipoprotein lipase gene. Association
with severity of coronary artery disease and diabetes. Circulation 1996; 93: 1339 ± 1345.
11 Palou A, Pico C, Bonet ML, Oliver P. Molecules in
focus: Uncoupling protein. Int J Biochem Cell Biol 1998;
30: 7 ± 11.
12 Anderson KM, Kannel WB. Obesity and disease. In BjoÈrntorp
P, Brodoff BN (eds). Obesity. Lippincott: Philadelphia, PA,
1992, pp 465 ± 473.
13 Mahley RW, Palaoglu KE, Atak Z, Dawson-Pepin J, Langlois
AM, Cheung V, Onat H, Fulks P, Mahley LL, Vakar F,
OzbayrakcËI S, GoÈkdemir O, Winkler W. Turkish heart
study: Lipids, lipoproteins, and apolipoproteins. J Lipid Res
1995; 36: 839 ± 859.
14 National Diabetes Data group. Classi®cation of diabetes mellitus and other categories of glucose intolerance. Diabetes
1979; 28: 1039 ± 1057.
15 Trinder P. Determination of glucose in blood using glucose
oxidase with an alternative oxygen acceptor. Ann Clin Biochem 1969; 69: 24 ± 27.
16 Barham D, Trinder P. An improved colour reagent for the
determination of blood glucose by the oxidase system. Analyst
1972; 97: 142 ± 145.
17 Allain CC, Poon LS, Chan CS, Richmond W, Fu PC. Enzymatic determination of total serum cholesterol. Clin Chem
1974; 20: 470 ± 475.
18 Bucolo G, David H. Quantitative determination of serum
triglycerides by the use of enzymes. Clin Chem 1973; 19:
476 ± 482.
19 Megraw RE, Dunn DE, Biggs HG. Manual and continuous¯ow colorimetry of triacylglycerols by a fully enzymic
method. Clin Chem 1979; 25: 273 ± 278.
20 Sambrook J, Fritsch EF, Maniatis T. In Molecular cloning,
2nd ed. Cold Spring Harbor Laboratory Press: Cold Spring
Harbor, 1989, pp 9.14 ± 9.23.
21 Cassard-doulcier AM, Bouillaud F, Chagnon M, Gelly C,
Dionne FT, Oppert JM, Bouchard C, Chagnon Y, Ricquier
D. The BclI polymorphism of the human uncoupling protein
(UCP) gene is due to a point mutation in the 50 ± ¯anking
region. Int J Obes 1996; 20: 278 ± 279.
22 Ahn YI, Kamboh MI, Hamman RF, Cole SA, Ferrell RE. Two
DNA polymorphisms in the lipoprotein lipase gene and their
associations with factors related to cardiovascular disease. J
Lipid Res 1993; 34: 421 ± 428.
23 Togan I, ErguÈven A. Red cell enzyme and blood
group polymorphisms in Turkey. Gene Geog 1994; 8: 75 ±
80.
24 Gerdes C, Gerdes LU, Hansen PS, Faergeman O. Polymorphisms in the lipoprotein lipase gene and their associations with
plasma lipid concentrations in 40-year-old Danish men.
Circulation 1995; 92: 1765 ± 1769.
25 Jemaa R, Tuzet S, Portos C, Betoulle D, Apfelbaum M,
Fumeron F. Lipoprotein lipase gene polymorphisms: associations with hypertriglyceridemia and body mass index in obese
people. Int J Obes 1995; 19: 270 ± 274.
26 Lonnqvist F, Thorne A, Nilsell K, Hoffstedt J, Arner P. A
pathogenic role of visceral fat b3-adrenoceptors in obesity. J
Clin Invest 1995; 95: 1109 ± 1116.
27 Lonnqvist F, Thorne A, Large V, Arner P. Sex differences
in visceral fat lipolysis and metabolic complications of obesity. Arteriorscler. Thromb. Vasc. Biol. 1997; 17(7): 1472 ±
1480.
28 Willett WC, Manson JE, Stampfer MJ, Colditz GA, Rosner B,
Speizer FE, Hennekens CH. Weight, weight change and
coronary heart disease in women: Risk within the normal
weight range. JAMA 1995; 273(6): 461 ± 465.
29 Borkan GA, Sparrow D, Wisniewski C, Vokonas PS.
Body weight and coronary disease risk: Patterns of risk
factor change associated with long-term weight change. The
Normative Ageing Study. Am J Epidemiol 1986; 124: 410 ±
419.
30 Anderson KM, Wilson PW, Garrison RJ, Castelli WP. Longitudinal and secular trends in lipoprotein cholesterol measurements in a general population sample. Atherosclerosis 1987;
68: 59 ± 66.
99
International Journal of Obesity
Genetic variants and metabolic disorders in obesity
AM Proenza et al
100
31 Lapidus L, Bengtsson C, Larsson B, Pennert K, Rybo E,
Sjostrom L. Distribution of adipose tissue and risk of cardiovascular disease and death: a 12-year follow-up of participants
in the population study of women in Gothenburg, Sweden. Br
Med J 1984; 289: 1257 ± 1261.
32 Caro J. Insulin resistance in obese and non obese man. J Clin
Endocrinol Metab 1990; 73: 691 ± 695.
International Journal of Obesity
33 Mitchell RJ, Earl L, Bray P, Fripp YJ, Williams J. DNA
polymorphisms at the lipoprotein lipase gene and their association with quantitative variation in plasma high-density lipoproteins and triacylglycerides. Human Biol 1994; 66: 383 ± 397.
34 Despres JP, Moorjani S, Lupien P-J, Tremblay A, Nadeau A,
Bouchard C. Genetic aspects of susceptibility to obesity and
related dyslipidemias. Mol Cell Biochem 1992; 113: 151 ± 169.