PAPER Interactions among the glucocorticoid receptor ... - Nature

112 downloads 0 Views 173KB Size Report
Interactions among the glucocorticoid receptor, lipoprotein lipase and adrenergic receptor genes and abdominal fat in the Que´bec Family Study. O Ukkola1,2, L ...
International Journal of Obesity (2001) 25, 1332–1339 ß 2001 Nature Publishing Group All rights reserved 0307–0565/01 $15.00 www.nature.com/ijo

PAPER Interactions among the glucocorticoid receptor, lipoprotein lipase and adrenergic receptor genes and abdominal fat in the Que´bec Family Study O Ukkola1,2, L Pe´russe3, YC Chagnon3, J-P Despre´s4 and C Bouchard1* 1

Pennington Biomedical Research Center, Baton Rouge, Louisiana, USA; 2Department of Internal Medicine, University of Oulu, Oulu, Finland; 3Physical Activity Sciences Laboratory, Division of Kinesiology, Laval University, Ste-Foy, Que´bec, Canada; and 4 Lipid Research Center, Laval University, Ste-Foy, Que´bec, Canada OBJECTIVE: To investigate whether interactions between glucocorticoid receptor (GRL), lipoprotein lipase (LPL) and adrenergic receptor (ADR) gene markers contribute to individual differences in indicators of adiposity and abdominal obesity, including visceral fat level. DESIGN AND SUBJECTS: Cross-sectional study; 742 individuals from the phase 2 of the Que´bec Family Study cohort. MEASUREMENTS: Total body fat assessed by hydrodensitometry and the sum of six skinfolds. Abdominal fat areas measured by computed tomography and adjusted for age, sex and total fat mass in all analyses. GRL Bcl I, a2A-ADR Dra I and b2-ADR Ban I markers were typed by Southern blot, and other markers by polymerase chain reaction technique. RESULTS: It is confirmed that the 4.5 kb allele of the GRL BclI polymorphism is associated with a higher amount of abdominal visceral fat (AVF) depot (P for trend < 0.001) independent of the level of total body fat. Furthermore, the a2-ADR Dra I variant is associated with lower cross-sectional areas of abdominal total (P ¼ 0.003) and subcutaneous (P ¼ 0.012) adipose tissue. Gene – gene interactions between GRL and a2-ADR genes affecting overall adiposity (P ¼ 0.016) as well as between GRL and b2-ADR genes (P ¼ 0.049) having influence on total abdominal fat levels were observed. When the three genes were considered together in the same analysis, significant interactions having influence on overall adiposity (P ¼ 0.017), abdominal total (P ¼ 0.032) and visceral fat (P ¼ 0.002) were observed. About 1 – 2% of the total variation in total fatness and abdominal fat was explained by these gene – gene interactions. CONCLUSION: There is an association between the GRL BclI polymorphism and increased AVF levels independent of the level of total body fat. The a2-ADR DraI variant is associated with a lower cross-sectional area of abdominal total fat. Numerous interactions between GRL and ADR markers on overall adiposity and total abdominal fat as well as between GRL, LPL and ADR genes on overall adiposity, abdominal total and visceral fat suggest that the genetic architecture of body fat content and adipose tissue distribution is complex although some genes, like GRL, may have ubiquitous effects. International Journal of Obesity (2001) 25, 1332 – 1339 Keywords: gene – gene interactions; abdominal fat

Introduction Factors that control the storage and mobilization of triglycerides in adipocytes are important regulators of fat accumulation in various adipose areas.1 The role of the adipose tissue lipoprotein lipase (LPL) in the regulation of uptake of triglycerides into adipose tissue has long been recognized.2

*Correspondence: C Bouchard, Pennington Biomedical Research Center, 6400 Perkins Road, Baton Rouge, LA 70808-4124, USA. E-mail: [email protected] Received 15 February 2000; revised 25 August 2000; accepted 27 September 2000

The adrenergic system modulates the breakdown of triglycerides in adipocytes through adrenoceptors (ADR).3 Glucocorticoids regulate both LPL activity and lipolysis4 and their effects are mediated through a specific glucocorticoid receptor (GRL).5 Thus, the GRL, LPL and ADR genes are relevant candidate genes for obesity and regional fat distribution phenotypes. However, earlier studies have failed to consistently demonstrate any major effect of these genes on regional fat distribution.6,7 One possible explanation is that the contribution of these genes is primarily through gene – gene interactions, which are likely to be ubiquitous in the genetic determinants of obesity.8 This is a topic that has

Gene interactions and obesity phenotypes O Ukkola et al

1333 received little attention thus far. Hence interactions between GRL, LPL and ADR genes could have effects on adiposity or regional fat distribution phenotypes that are not observed when the gene effects are analyzed individually. For example, the GRL9 and b2-ADR10 genes are in close proximity on chromosome 5, which could have favored similar evolutionary features and perhaps functional interactions. We therefore investigated the evidence for interactions between polymorphisms in GRL, LPL and ADR genes on adiposity and regional fat distribution.

Material and methods The Que´bec Family Study (QFS) cohort has been described previously.11 In the present study, based on phase 2 of the QFS project, 742 individuals from French-Canadian families living in and around Que´bec city are included. Informed written consent was obtained from all subjects and the QFS project had been approved by the Medical Ethics Committee of Laval University.

Phenotype measurements Body mass index (BMI) was calculated as body weight (kg) divided by height squared (m2). Body density obtained by underwater weighting12 was converted to percentage body fat using the equation of Siri13 with residual pulmonary volume measured with the helium dilution method.14 Fat mass (FM) and fat-free mass (FFM) were obtained from percentage body fat and body weight. Skinfolds were measured on the left side of the body according to the recommendations of the International Biological Programme15 and included six sites: biceps, triceps, medial calf, subscapular, suprailiac and abdominal. The sum of the six skinfold thickness measures was considered as an indicator of total subcutaneous fat. Total abdominal and visceral fat areas were obtained using computed tomography as described earlier.16

DNA analyses Genomic DNA was isolated from lymphoblastoid cell cultures17 by digestion with proteinase K and extraction with phenol chloroform. Genotyping of the b2- and b3-ADR markers. The T to C transition at codon 64 in the b3-ADR gene leads to a replacement of tryptophan by arginine and generates a new MspI restriction site. Specific primers covering this MspI restriction site were generated as reported earlier.6 Each 10 ml reaction contained 250 ng of genomic DNA, 0.3 mM of each primer, 0.2 mM of each of the dNTPs and 2.5 units Taq DNA polymerase in a standard buffer and Q solution (Qiagen Inc., Chatsworth, CA, USA). The PCR reaction solutions were incubated at 94 C for 3 min, 50 C for 1 min, 72 C for 1 min followed by 35 cycles at 94 C for 30 s, 60 C for 30 s, 72 C for 45 s and one cycle at 72 C for 10 min

using a thermal cycler (model 9600; Perkin-Elmer Cetus Instruments, Branchburg, NJ). PCR products were digested by adding 7.5 U of MspI enzyme for 18 h at 37 C and the fragments obtained were separated on a 3% agarose gel and visualized under UV light after staining with ethidium bromide. The manufacturer of the restriction enzymes is New England Biolabs (Beverly, MA, USA) unless specified otherwise. PCR analysis of the Gln27Glu polymorphism in codon 27 of the ADR b2 gene was carried out in a volume of 20 ml containing 150 ng DNA, 0.3 mM of each primer, 0.2 mM of each of the dNTPs, 1.0 unit Taq polymerase and 1standard buffer plus 10% DMSO. The primers were those reported earlier.18 The PCR was started at 95 C for 3 min, 60 C for 1 min and 72 C for 1 min followed by 35 cycles at 95 C for 30 s, 60 C for 30 s, 72 C for 45 s and one cycle at 72 C for 10 min. The amplified product was digested at 37 C for 18 h with 1 unit of ItaI (Roche Diagnostics Corporation, Indianapolis, IN, USA). The fragments were separated on 2.0% agarose gel. The PCR analysis for the Arg16Gly polymorphism in codon 16 of the ADR b2 gene was carried out otherwise as for the Gln27Glu polymorphism but without using DMSO, and the annealing temperature was 57 C. The primers were those reported earlier.18 The amplified product was digested at 60 C for 4.5 h with 4 U of BsrDI. The fragments were separated on 3.0% agarose gel. Genotyping of the LPL marker. A Ser447Ter mutation in exon 9 of the LPL gene caused by a C-G transversion results in a premature termination codon.19 This leads to a truncated protein lacking the two carboxyl-terminal amino acids (SER-GLY). Modified primers covering this site and creating a HinfI restriction site in the presence of the G allele were those reported earlier by Garenc et al.20 In the PCR studies of the Ser447Ter polymorphism, a volume of 20 ml containing 250 ng DNA, 0.45 mM of each primer, 0.3 mM of each of the dNTPs, 1.1 unit Taq polymerase, 1.9 mM MgCl2 and standard buffer plus Q-solution was used. The PCR was started at 94 C for 3 min, 53 C for 1 min and  72 C for 1 min followed by 40 cycles at 94 C for 30 s, 53 C for 30 s and 72 C for 45 s and one cycle at 72 C for 10 min. The amplified product was digested at 37 C for at least 18 h with 10 U of HinfI. The fragments were separated on 3.0% agarose gel. Southern blot analysis. All of the samples of the a2A-ADR DraI, b2-ADR BanI and GRl BclI polymorphisms were identified by Southern blot technique. The human a2- and b2-ADR genomic probes were obtained from the American Type Culture Collection (ATCC, Maryland, USA). The human GRL genomic DNA was provided by R Evans21 and the human LPL cDNA clone by R Lawn.22 Five micrograms of genomic DNA were digested for 18 – 20 h with 30 U of the DraI for a2-ADR, BanI for b2-ADR and BclI for GRL according International Journal of Obesity

Gene interactions and obesity phenotypes O Ukkola et al

1334 to conditions recommended by the manufacturer. Southern blot analysis was performed as described earlier.7

performed quite well. The only negative impact was a small reduction in power. The standard errors were slightly enlarged but, most importantly, type I error was unaffected. Given this, we do not believe that the dependencies or relatedness of the subjects in families causes any real problems in this type of analysis.

Statistical analyses All the analyses were performed with the SAS Statistical Software (SAS Institute Inc., Cary, NC). A chi-square test was performed to assess whether the observed genotype frequencies were in a Hardy – Weinberg equilibrium. Associations between the gene markers and phenotypes were tested by ANCOVA using the general linear model (GLM) procedure. BMI, FM, FFM and sum of six skinfolds were adjusted for age and sex. Abdominal fat areas were adjusted for total FM, age and sex. Data adjusted in this fashion are independent of the overall level of adiposity and reflect predominantly fat accumulation in the abdominal area. Gene – gene interactions were tested with the GLM procedure by including the main gene effects, interaction terms and covariates in the same model. All the family members were included in the analyses. Although it is commonly believed that the relatedness of the subjects in family study cohorts may cause problems in association analyses (more false positives), a recent simulation study (M Province, T Rice, DC Rao, Washington University, unpublished data) suggests that this is not the case. In that study, the data were analyzed by four methods, where the least squares method used in the present report was one of them; the other three methods treated dependencies in different ways. Results showed that failure to incorporate dependencies did not induce any bias. For moderate familial correlations, as seen in most family studies (including the current one), ignoring the dependencies by using ANOVA Table 2

Results The phenotypic characteristics of the 742 subjects are shown in Table 1. Men were heavier, had lower FM and higher FFM, lower subcutaneous fat but higher abdominal visceral fat (AVF) compared to women (Table 1). Independent and interactive effects of each polymorphism on regional fat distribution and their contributions to the phenotypes variance are shown in Table 2. Abdominal fat

Table 1 Characteristics of the subjects of the Que´bec Family Study sample (n ¼ 742). Variables

Male

Age (y) 42.3  0.9 Body weight (kg) 81.1  1.1 2 27.2  0.3 Body mass index (kg=m ) Fat mass (kg) 19.5  0.7 Fat-free mass (kg) 60.9  0.5 Subcutaneous fat (mm) (sum of 96.1  3.0 six skinfolds) 2 341.5  12.9 Abdominal total fat area (cm ) 2 Abdominal visceral fat area (cm ) 126.2  5.5 215.2  8.8 Abdominal subcutaneous fat area (cm2 )

Female

n

322 42.7  0.8 320 69.5  1.0* 320 27.3  0.4 282 23.4  0.8* 282 45.6  0.4* 312 143.0  3.6*

Values are means  s.e.m. P < 0.0005 between males and females.

Abdominal fat Total fat

r

2

Total

r

2

Subcutaneous

r

2

Visceral

Independent effects a2-ADR DraI b2-ADR BanI b2-ADR Gln27Glu b2-ADR Arg16Gly b3-ADR Trp64Arg LPL Ser447Ter GRL BclI

No No No No No No Yes

— — — — — — 0.020

Yes No No No No No Yes

0.002 — — — — — 0.002

Yes No No No No No Yes

0.002 — — — — — 0.002

No No No No No No Yes

— — — — — — 0.011

Interactive effects GRLa2-ADR a GRLb2-ADR GRLb3-ADR LPLa2-ADR LPLb2-ADR LPLb3-ADR GRLLPLa2-ADR b GRLLPLb2-ADR GRLLPLb3-ADR

Yes No No No No No Yes No No

0.007 — — — — — 0.015 — —

No Yes No No No No No Yes No

— 0.001 — — — — — 0.002 —

No No No No No No No No No

— — — — — — — — —

No No No No No No Yes No No

— — — — — — 0.011 — —

a

b

BanI; Gln27Glu markers. Yes ¼ significant association (P < 0.05). No ¼ nonsignificant association (P > 0.05).

International Journal of Obesity

420 416 416 337 337 389

222 424.3  13.4* 293 222 97.4  3.8* 293 222 326.9  10.6* 293

Independent and interactive effects of each polymorphism on total fatness and abdominal fat

Polymorphisms

n

2

r

Gene interactions and obesity phenotypes O Ukkola et al

1335 areas are all adjusted for age, sex and total fat mass in all these analyses.

Independent effects of each polymorphism As reported earlier,23 the homozygotes for the 6.3 kb allele of the DraI polymorphism in the a2-ADR gene had the lowest mean abdominal subcutaneous (P ¼ 0.012) and total fat areas (P ¼ 0.003), but the DraI variant was not associated with overall adiposity. The b2-ADR, b3-ADR or LPL markers were not associated with overall adiposity nor with abdominal fat phenotypes (Table 2). The 4.5=2.3 kb heterozygotes of the GRL BclI polymorphism had higher BMI (P for trend 0.012), FFM (P for trend 0.003), FM (P for trend 0.002), and total subcutaneous fat (P for trend < 0.001; Table 3). The homozygotes for the 4.5 kb allele had higher abdominal total fat (P for trend 0.005) compared to heterozygotes or homozygotes for the 2.3 kb allele (Figure 1), while the homozygotes or heterozygotes for the 4.5 kb allele had higher amount of AVF compared to the homozygotes for the 2.3 kb allele (P for trend < 0.001).

Gene – gene interactions Between GRL and a2-ADR genes. A significant interaction was observed between the GRL and a2-ADR gene markers for total subcutaneous fat (P for interaction 0.016), with noncarriers of 4.5 kb allele of the GRL BclI and carriers of the 6.3 kb allele of the a2-ADR DraI polymorphisms having the lowest amount (97.3  6.7; n ¼ 87) compared to carriers (127.8  3.6; n ¼ 306; P < 0.001) or non-carriers (117.8  4.6; n ¼ 185; P ¼ 0.011) of the 4.5 kb allele of the BclI and noncarriers of the 6.3 kb allele of the DraI polymorphisms. Between GRL and b2-ADR genes. The interaction between GRL BclI and b2-ADR BanI polymorphisms was significant for total abdominal fat area (adjusted for age, sex and total FM) in the whole cohort (Figure 2; P for interaction 0.049). The carriers of the 4.5 kb allele of the GRL BclI and 3.4 kb allele of the b2-ADR BanI polymorphisms had the highest mean abdominal total fat area. Two other b2-ADR markers,

Gln27Glu and Arg16Gly, which are in strong linkage disequilibrium with the BanI marker (w2 from 122.74 to 441.08, P < 0.001), did not show any interactions with the GRL marker. Between GRL, LPL and a2-ADR genes. When the three (GRL, LPL and a2-ADR) genes were considered in the same analysis, there were interactions among them affecting total subcutaneous fat (P for interaction 0.017; Figure 3). Subjects who were carriers of the Ter allele of the LPL Ser447Ter polymorphism and non-carriers of the BclI 4.5 kb allele plus carriers of the DraI 6.3 kb allele had the lowest values, whereas subjects who were non-carriers of the Ter allele and carriers of the 4.5 kb allele plus carriers of the 6.3 kb allele had the highest values. Significant interactions were observed for AVF (P ¼ 0.002) between the GRL, LPL and a2-ADR genes (Figure 4). The mean level of AVF (adjusted for age, sex and total fat mass) changed in a stepwise manner. Carriers of the 4.5 kb allele of the GRL BclI polymorphism and non-carriers of the 6.3 kb allele of the a2-ADR DraI polymorphism had the highest amount, while non-carriers of the 4.5 kb allele of the GRL BclI polymorphism and carriers of the 6.3 kb allele of the a2ADR DraI polymorphism had the lowest amount. This was observed among the non-carriers of the Ter447 allele of the LPL Ser447Ter polymorphism (Figure 4). Among the carriers of the Ter447 allele of the LPL Ser447Ter polymorphism, the amount of AVF was highest in the carriers of the 4.5 kb allele of the GRL BclI polymorphism and 6.3 kb allele of the a2ADR DraI polymorphism (Figure 4). Between GRL, LPL and b2-ADR genes. The subjects who were non-carriers of the Ter allele of the LPL Ser447Ter polymorphism and also 4.5 kb and Glu allele carriers of the GRL BclI and b2-ADR Gln27Glu variants, respectively, had the highest amount of abdominal total fat (P for interaction 0.032; data not shown). The contribution of the gene – gene interactions to the phenotypes variance. The GRL polymorphism explained 1.1% of the total variation in AVF beyond the covariate effects.

Table 3 Characteristics of the subjects by glucocorticoid receptor BclI genotype Genotype

Number of subjects Body mass index (kg=m2) Fat mass (kg) Fat free mass (kg) Sum of 6 skinfolds (mm)

4.5=4.5 kb Group 1

4.5=2.3 kb Group 2

2.3=2.3 kb Group 3

P for trend

90 26.3  0.8 20.6  1.4 51.6  0.8 119.2  6.6

358 28.1  0.4* 23.4  0.7** 53.6  0.4*** { 131.4  3.4

288 26.5  0.4 19.6  0.8 51.5  0.5 111.3  3.8

0.012 0.002 0.003 < 0.001

Values are means  s.e.m. *P ¼ 0.038 between 2 and 1, P ¼ 0.008 between 2 and 3; **P ¼ 0.001 between 2 and 3; ***P ¼ 0.027 between 2 and 1, { P ¼ 0.002 between 2 and 3; P < 0.001 between 2 and 3.

International Journal of Obesity

Gene interactions and obesity phenotypes O Ukkola et al

1336 The contributions of covariates to the AVF variance were as follows: sex 8.6%, age 14.1% and FM 33.5%. The variance explained by all these reached 70%.

The GRL and a2-ADR interaction explained 0.7% of the variation in total subcutaneous fat (Table 2). About 1.5% of the total variation in subcutaneous fat levels and 1.1% of the total variation in AVF were explained by the GRL, LPL and a2-ADR markers interactions. The independent effect of each genetic marker persisted when the interaction term was added to the model but the effect of the GRL variant on abdominal total fat disappeared.

Discussion

Figure 1 Mean abdominal subcutaneous (upper panel), visceral (middle panel) and total (lower panel) fat by glucocorticoid receptor (GRL) BclI genotype in the whole group (mean  s.e.m.). Adjusted for age, sex and total fat mass. P for trend ¼ 0.043 for the upper panel, < 0.001 for the middle panel and 0.005 for the lower panel. Number of subjects is embedded in the columns. *P ¼ 0.015 between 4.5=4.5 kb and 4.5=2.3 kb alleles; **P ¼ 0.001 between 4.5=4.5 kb and 2.3=2.3 kb; ***P < 0.001 between 4.5=2.3 kb and 2.3=2.3 kb; {P ¼ 0.008 between 4.5=4.5 kb and 4.5=2.3 kb, P ¼ 0.001 between 4.5=4.5 kb and 2.3=2.3 kb.

International Journal of Obesity

Our data demonstrate that the a2-ADR DraI variant is associated with abdominal total and subcutaneous fat areas and the GRL BclI variant with abdominal total, subcutaneous and visceral fat areas independent of the level of total body fat. In addition, the results suggest that there are gene – gene interactions between GRL and a2-ADR genes affecting overall adiposity as well as between GRL and b2-ADR genes on total abdominal fat levels. When the three candidate genes (GRL, LPL and ADR) were considered together in the same analysis, significant interactions having influence on overall adiposity, abdominal total and AVF levels were observed. About 1 – 2% of the total variation in total fatness and abdominal fat were explained by these gene – gene interactions. The findings on AVF are in accordance with our previous report in which an association of the GRL BclI polymorphism with AVF was seen in lean subjects.7 The present study, based on a larger QFS sample, shows that the association of the 4.5 kb allele with a higher AVF area is true for any level of body FM since AVF data were adjusted for the latter. The association of the GRL variant with AVF is therefore independent of total FM and it explains why the heterozygotes who had the highest BMI and FM did not present the highest accumulation of AVF. Thus the effect of GRL BclI is on AVF level per se and not on total adiposity. In earlier studies, associations between GRL variants and hyperinsulinemia24 have been reported. However, two linkage studies with obesity phenotypes have given negative results.25,26 How could the GRL marker be associated with the level of AVF? Glucocorticoids, in the presence of insulin, stimulate adipose tissue LPL activity and are also involved in the regulation of lipolysis.4 Their effects are mediated via specific glucocorticoid receptors.5 In addition, it has been shown that glucocorticoid receptor density is higher in visceral than in subcutaneous adipocytes.27 The GRL BclI polymorphism might enhance the effects of glucocorticoids in adipose depots that are more sensitive to them even under physiological conditions. The latter could lead to an increased storage of triglycerides and accumulation of fat into the abdominal visceral depot. A recent report by Panarelli et al28 suggests that the BclI polymorphism may lead to tissue-specific differences in GRL expression by affecting the GRL promoter region and to enhanced sensitivity to glucocorticoids in selected tissues. A linkage disequilibrium with another mutation inside or outside the GRL locus

Gene interactions and obesity phenotypes O Ukkola et al

1337

Figure 2 Mean abdominal total fat in relation to the GRL and b2-adrenergic receptor (ADR) genotypes in the whole group. Adjusted for age, sex and total FM. For the GRLb2-ADR marker interaction, P ¼ 0.049. *P ¼ 0.007 between groups 3 and 4.

Figure 3 Mean total subcutaneous fat (the sum of the six skinfolds) in relation to the LPL, GRL and a2-ADR genotypes in the whole group. For the LPLGRLa2-ADR marker interaction, P ¼ 0.017. *P ¼ 0.005 between 2 and 3; **P ¼ 0.005 between 4 and 1, P < 0.001 between 4 and 2, P ¼ 0.010 between 4 and 5; ***P ¼ 0.001 between 8 and 1, P < 0.001 between 8 and 2, P ¼ 0.018 between 8 and 3, P ¼ 0.002 between 8 and 5, P ¼ 0.024 between 8 and 6, P ¼ 0.008 between 8 and 7.

International Journal of Obesity

Gene interactions and obesity phenotypes O Ukkola et al

1338

Figure 4 Mean AVF in relation to the LPL, GRL and a2-ADR genotypes in the whole group. Adjusted for age, sex and total FM. For the LPLGRLa2-ADR marker interaction, P ¼ 0.002. *P ¼ 0.034 between 1 and 2, P ¼ 0.001 between 1 and 3, P < 0.001 between 1 and 4, P < 0.001 between 1 and 5, P ¼ 0.026 between 1 and 7; **P ¼ 0.025 between 2 and 4; ***P ¼ 0.023 between 6 and 3, P ¼ 0.001 between 6 and 4, P ¼ 0.003 between 6 and 5.

influencing steroid metabolism has been postulated7,28 and could also contribute to the observed associations. In addition to the effects of the GRL locus, the data of the present study show that several interactions between the GRL, LPL and ADR genes are associated with variations in regional fat distribution phenotypes. Gene – gene interactions involving genes in close proximity on the same chromosome are of particular interest, even though the physical proximity is not a prerequisite for such interactions to occur. In this regard, GRL9 and b2-ADR10 genes are located in close proximity on chromosome 5. Although Takami et al26 failed to find any linkage of GRL or b2-ADR markers with blood pressure or BMI, significant interaction effects between DNA sequence variants in these two genes on obesity and its comorbidities could exist. Indeed, the data of the present study show that the interactions between GRL and b2-ADR markers were associated with variation in abdominal total fat level. In fact, the influence of the 4.5 kb allele of GRL BclI polymorphism on abdominal fat was seen only in carriers of the 3.4 kb allele of the b2-ADR BanI polymorphism. Although the exact mechanisms remain to be determined, an interaction between glucocorticoids and sympathetic nervous system genes affecting adipocyte metabolism, particularly lipolysis, is an attractive hypothesis. The functional significance of the BanI polymorphism remains unknown. Its effects may be mediated through other b2-ADR markers which are in strong linkage disequilibrium with it29 International Journal of Obesity

and which have earlier been reported to have functional properties.30 The fact that we observed interaction effects between GRL, LPL and ADR gene variants influencing overall adiposity, abdominal total and visceral fat levels is also of interest. One can speculate that the increase in the amount of adiposity is seen among those subjects who are jointly exposed to the potentially deleterious metabolic effects of the three variants in GRL, LPL and ADR systems. Thus, functional balance between storage and release of triglycerides may change because of interactions between the GRL and ADR variants affecting glucocorticoid sensitivity and lipolytic activity, respectively. This balance, which, among other factors, determines the amount of lipid stored within different adipose regions,1 depends also on the DNA sequence characteristics at the LPL locus. In conclusion, there is an association between the GRL BclI polymorphism and the AVF depot independent of the level of total body fat. However, age, sex and FM explain most of the variance in AVF levels and the BclI variant contributes only marginally. In addition, the a2-ADR DraI variant is associated with lower amount of abdominal fat. Numerous interactions between the GRL and ADR genes on overall adiposity and total abdominal fat, as well as between GRL, LPL and ADR genes on overall adiposity, abdominal total and visceral fat, suggest that the genetic architecture of body fat content and adipose

Gene interactions and obesity phenotypes O Ukkola et al

tissue distribution is complex although some genes, like GRL, may have ubiquitous effects.

Acknowledgements This work was supported by grants from the Medical Research Council of Canada (PG-11811, MT-13960, GR15187), the Finnish Cultural Foundation and Medical Council of the Academy of Finland, and the Pennington Biomedical Research Center. We thank Monique Chagnon, ART, for her technical assistance. C Bouchard is supported in part by the George A Bray Chair in Nutrition. References 1 Bouchard C, Despre´s J-P, Maurie´ge P. Genetic and nongenetic determinants of regional fat distribution. Endocr Rev 1993; 14: 72 – 93. 2 Eckel RH. Lipoprotein lipase. A multifunctional enzyme relevant to common metabolic diseases. New Engl J Med 1989; 320: 1060 – 1068. 3 Fain JN, Garcia-Sainz JA. Adrenergic regulation of adipocyte metabolism. J Lipid Res 1983; 24: 945 – 966. 4 Cigolini M, Smith U. Human adipose tissue in culture. VIII. Studies on the insulin-antagonistic effects of glucocorticoids. Metabolism 1979; 28: 502 – 510. 5 Bamberger CM, Schulte HM, Chrousos GP. Molecular determinants of glucocorticoid receptor function and tissue sensitivity to glucocorticoids. Endocr Rev 1996; 17: 245 – 261. 6 Gagnon J, Maurie´ge P, Roy S, Sjostrom D, Chagnon YC, Dionne FT, Oppert JM, Pe´russe L, Sjostrom L, Bouchard C. The Trp64Arg mutation of the b3 adrenergic receptor gene has no effect on obesity phenotypes in the Que´bec Family Study and Swedish Obese Subjects cohorts. J Clin Invest 1996; 98: 2086 – 2093. 7 Buemann B, Vohl M-C, Chagnon M, Chagnon YC, Gagnon J, Pe´russe L, Dionne F, Despre´s J-P, Tremblay A, Nadeau A, Bouchard C. Abdominal visceral fat is associated with a Bcl I restriction fragment length polymorphism at the glucocorticoid receptor gene locus. Obes Res 1997; 5: 186 – 192. 8 Bouchard C. Genetics of obesity: overview and research directions. In: Bouchard C (ed). The genetics of obesity. CRC Press: Boca Raton, FL; 1994. pp 223 – 233. 9 Theriault A, Boyd E, Harrap SB, Hollenberg SM, Connor JM. Regional chromosomal assignment of the human glucocorticoid receptor gene to 5q31. Hum Genet 1989; 83: 289 – 291. 10 Kobilka BK, Frielle T, Dohlman HG, Bolanowski MA, Dixon RA, Keller P, Caron MG, Lefkowitz RJ. Delineation of the intronless nature of the genes for the human and hamster b2-adrenergic receptor and their putative promoter region. J Biol Chem 1987; 262: 7321 – 7327. 11 Bouchard C. Genetic epidemiology, association and sib-pair linkage: results from the Que´bec Family Study. In: Bray GA, Ryan DH (eds). Molecular and genetic aspects of obesity, Vol 5. Pennington Center Nutrition Series, Louisiana State University Press: Baton Rouge, LA; 1996. pp 470 – 481. 12 Behnke AR, Wilmore JH (eds). Evaluation and regulation of body build and composition. Prentice-Hall: Englewood Cliffs, NJ; 1974. pp 20 – 37. 13 Siri WE. The gross composition of the body. In: Lawrence JH, Tobias CA (eds). Advances in biological and medical physics, Vol 4. Academic Press: NewYork; 1956. pp 239 – 280. 14 Meneely GR, Kaltreider NL. The volume of the lung determined by helium dilution: description of the method and comparison with other procedures. J Clin Invest 1949; 28: 129 – 139.

15 Weiner JS, Lourie JA (eds). Human biology: a guide to field methods. Blackwell: Oxford; 1969. 16 Ferland M, Despre´s JP, Tremblay A, Pinault S, Nadeau A, Moorjani S, Lupien PJ, Theriault G, Bouchard C. Assesment of adipose tissue distribution by computed axial tomography in obese women: association with body density and anthropometric measurements. Br J Nutr 1989; 61: 139 – 148. 17 Neitzel H. A routine method for the establishment of permanent growing lymphoblastoid cell lines. Hum Genet 1986; 73: 320 – 326. 18 Large V, Hellstrom L, Reynisdottir S, Lonnqvist F, Eriksson P, Lannfelt L, Arner P. Human beta-2 adrenoceptor gene polymorphisms are highly frequent in obesity and associate with altered adipocyte beta-2 adrenoceptor function. J Clin Invest 1997; 100: 3005 – 3013. 19 Hata A, Robertson M, Emi M, Lalouel JM. Direct detection and automated sequencing of individual alleles after electrophoretic strand separation: identification of a common nonsense mutation in exon 9 of the human lipoprotein lipase gene. Nucleic Acids Res 1990; 18: 5407 – 5411. 20 Garenc C, Pe´russe L, Gagnon J, Chagnon YC, Bergeron J, Despre´s J-P, Province MA, Leon AS, Skinner JS, Wilmore JH, Rao DC, Bouchard C. Linkage and association studies of the lipoprotein lipase gene with post-heparin plasma lipase activities, body fat and plasma lipid and lipoprotein concentrations: the HERITAGE Family Study. Metabolism 2000; 49: 432 – 439. 21 Hollenberg SM, Weinberg C, Ong ES, Cerelli G, Oro A, Lebo R, Thompson EB, Rosenfed MG, Evans RM. Primary structure and expression of a functional human glucocorticoid receptor cDNA. Nature 1985; 318: 635 – 641. 22 Wion KL, Kirchgessner TG, Lusis AJ, Schotz MC, Lawn RM. Human lipoprotein lipase complementary DNA sequence. Science 1987; 235: 1638 – 1641. 23 Ukkola O, Rankinen T, Weisnagel SJ, Sun G, Pe´russe L, Chagnon YC, Despre´s J-P, Bouchard C. Interactions among the a2-, b2- and b3-adrenergic receptor genes and obesity-related phenotypes in the Quebec Family Study. Metabolism 2000; 49: 1063 – 1070. 24 Weaver JU, Hitman GA, Kopelman PG. An association between a BclI restriction fragment length polymorphism of the glucocorticoid receptor locus and hyperinsulinaemia in obese women. J Mol Endocrinol 1992; 9: 295 – 300. 25 Cle´ment K, Philippi A, Jury C, Pividal R, Hager J, Demenais F, Basdevant A, Guy-Grand B, Froguel P. Candidate gene approach of familial morbid obesity: linkage analysis of the glucocorticoid receptor gene. Int J Obes Relat Metab Disord 1996; 20: 507 – 512. 26 Takami S, Wong ZYH, Stebbing M, Harrap SB. Linkage analysis of glucocorticoid and b2-adrenergic receptor genes with blood pressure and body mass index. Am J Physiol 1999; 276: H1379 – H1384. 27 Rebuffe´-Scrive M, Lundholm K, Bjorntorp P. Glucocorticoid hormone binding to human adipose tissue. Eur J Clin Invest 1985; 15: 267 – 271. 28 Panarelli M, Holloway CD, Fraser R, Connell JMC, Ingram MC, Anderson NH, Kenyon CJ. Glucocorticoid receptor polymorphism, skin vasoconstriction, and other metabolic intermediate phenotypes in normal human subjects. J Clin Endocrinol Metab 1998; 83: 1846 – 1852. 29 Large V, Hellstrom L, Reynisdottir S, Lonnqvist F, Eriksson P, Lannfelt L, Arner P. Human beta-2 adrenoceptor gene polymorphisms are highly frequent in obesity and associate with altered adipocyte beta-2 adrenoceptor function. J Clin Invest 1997; 100: 3005 – 3013. 30 Green SA, Turki J, Innis M, Liggett SB. Aminoterminal polymorphisms of the human b2-adrenergic receptor impart distinct agonist-promoted regulatory properties. Biochemistry 1994; 33: 9414 – 9419.

1339

International Journal of Obesity