¹University Department of Clinical Biochemistry, Addenbrooke's Hospital, Cambridge, UK

²Department of Public Health and Primary Care, Institute of Public Health, University of Cambridge, UK

³Section of Endocrinology, Imperial College School of Medicine, London, UK

⁴International Diabetes Institute, Caulfield, Victoria, Australia

Correspondence to: DJ Halsall, Department of Clinical Biochemistry, Addenbrooke's Hospital, Cambridge, CB2 2QR, UK. E-mail: djh44@hermes.cam.ac.uk

OBJECTIVE: To investigate whether genetic variation at the UCP3 locus contributes to human obesity.

SUBJECTS: Ninety-one obese children (BMI>4 standard deviations from age related mean) and 419 Caucasian adults from the Isle of Ely Study.

DESIGN: Single strand conformation polymorphism (SSCP) analysis was used to scan the coding region of the UCP3 gene in 91 severely obese children. A common polymorphism identified in this gene (c-55t) has been shown to associate with lower UCP3 mRNA expression. Polymerase chain reaction-based forced restriction digestion was used to detect this allele in Caucasian adults. Multiple regression analysis was used to determine associations between the c-55t genotype and anthropometric, energetic and biochemical indices relevant to obesity.

MEASUREMENTS: For the obese children, SSCP analysis and sequencing of variants were carried out. For the Isle of Ely Study, c-55t genotype and anthropometric (body mass index, waist-hip ratio, percentage body fat), energetic (dietary fat intake, physical activity index, adjusted metabolic rate, maximum oxygen consumption) and biochemical indices (pre- and post-glucose challenge plasma triglycerides, non-esterified fatty acids, insulin and glucose) were determined.

RESULTS: A previously reported missense mutation (V102I) was detected in a single obese Afro-Carribean child. Twenty-one percent of the genes examined in the Isle of Ely study carried the c-55t promoter variant. Age-adjusted body mass index (BMI) was significantly (P=0.0037) lower in carriers of this variant.

CONCLUSION: Mutations in the coding sequence of UCP3 are unlikely to be a common monogenic cause of severe human obesity. In a Caucasian population the UCP3 c-55t polymorphism is negatively associated with BMI.

International Journal of Obesity (2001) 25, 472-477

uncoupling proteins; UCP3; obesity; BMI

Introduction

The gene encoding human uncoupling protein 3 (UCP3) was identified by virtue of its sequence homology to rat UCP1, a brown adipose tissue (BAT) specific protein responsible for the uncoupling of mitochondria. The protein product of UCP3 is expressed mainly in muscle that, due the relative absence of BAT in humans, is thought to be a major tissue involved in non-shivering thermogenesis. At present there is no direct evidence to suggest that UCP3 uncouples muscle mitochondria. Alternatively, its role may be the integration of carbohydrate and fat metabolism.¹ As UCP3 is likely to be involved in the regulation of energy balance it is a candidate gene for the pathogenesis of metabolic disorders such as obesity and diabetes.²

Several variant forms of the human UCP3 gene have been described. Two rare mutations^3,4 identified in obese subjects have been shown to impair UCP3 uncoupling activity in yeast expression studies,⁵ although the segregation of these mutations with obesity is not complete. Of the common polymorphisms identified in the coding sequence of UCP3, the majority have no obvious effect on biological phenotypes such as body composition, resting metabolic rate or biochemical studies of mitochondrial function.^6,7 However weak associations between silent polymorphisms in UCP3 and obesity or diabetes-related phenotypes have been reported.⁸ A common polymorphism six base pairs upstream of a putative TATA box transcription initiation site (c-55t) has been identified in UCP3.^9,10,11 mRNA expression is significantly higher in carriers of the rarer allele,¹⁰ even though the putative TATA box is unlikely to be that recognized in vivo.¹² No difference in c-55t genotype frequency was detected between lean and obese Pima Indians,¹⁰ but the polymorphism associates with waist-hip ratio in European and South Indian adult women.⁹

To further investigate whether variation in the UCP3 gene might play a role in the aetiology of human obesity two studies were undertaken. Firstly the entire coding sequence of UCP3 was screened for mutations in a cohort of subjects with severe early-onset obesity. Secondly, the effects of the previously reported c-55t polymorphism on anthropometric, energetic and biochemical indices in a large well-characterized Caucasian population-based study (The Isle of Ely Study)^13,14 were examined.

Subjects and methods

Subjects

Early-onset obesity: Ninety-one subjects with severe early-onset obesity were studied. None had a recognized clinical syndrome or a structural hypothalamic cause for their obesity. In all cases obesity was manifest before the age of 10 y and body mass index (BMI) in all subjects was>4 s.d. above the mean for their age.

Ely study: The methods used to obtain the data for the Isle of Ely Prospective Study have previously been described in detail.^13,14 Unrelated subjects for the study were recruited from a general practice register in Ely, Cambridgeshire, UK. Briefly, a letter of invitation was sent to a random selection of subjects not previously known to have diabetes between the ages of 40 and 65 y at time of recruitment to the base-line study in 1990-1992. After a 10 h fast, subjects attended a clinical examination that included a dietary and medical questionnaire, anthropometric and energetic measurements and an oral glucose tolerance test (OGTT). The analysis in this report is restricted to a random sample of 419 out of 1071 non-diabetic subjects who were genotyped at the UCP3 c-55t locus and who underwent a follow-up OGTT after a mean interval of 4.5 y.

Physical measurements: Fat percentage was measured using a Bodystat impedance monitor. Mass adjusted mean metabolic rate (AMR) was calculated from oxygen consumption measurement at rest. Physical activity level (PAL) and maximum oxygen consumption during exercise (VO_2max) were measured as described previously.^15,16

Dietary recall: Habitual diet during the previous year was assessed using a self-completion, semi-quantitative food frequency questionnaire. Grams of each food consumed per day were calculated and converted to nutrient intakes by referring to food tables. From these nutrient intakes, the fat intake percentage in the diet was derived.

Oral glucose tolerance test and biochemical analyses: The OGTT consisted of a 75 g oral glucose challenge with collection of venous blood samples at baseline, 30 min and 120 min. Samples of sera and plasma were immediately separated, kept on ice and stored at -70°C within 4 h. Biochemical analyses were performed as described previously.^14,17 The 30 min insulin increment was included in the analysis as a surrogate measure for insulin secretion following the OGTT and was calculated by dividing the difference between 30 min insulin and fasting insulin concentrations by the 30 min glucose concentration.¹⁸ The area under the curve (AUC) for plasma NEFA was used as a composite measure of NEFA suppression following the OGTT.¹⁹

Methods

DNA extraction protocol: Whole blood was collected into EDTA tubes (Sarstedt, Leicester UK). A total of 419 random subjects were taken from the Ely study for genotyping. The blood was centrifuged; plasma was removed and around 4 ml of packed cells were frozen at -20°C. DNA was extracted from thawed cells as described previously.²⁰

SSCP and DNA sequencing protocol: Single-strand conformation polymorphism (SSCP) analysis was performed as described previously²¹ using published oligonucleotide primers.⁷ SSCP variants were sequenced in both directions with the same oligonucleotide primers using Perkin Elmer/ABI Taq FS Big Dye cycle sequencing reagents and an ABI 310 automated fluorescent DNA sequencer (Warrington, UK).

UCP3 genotyping: A total of 419 individuals were genotyped for the c-55t polymorphism using a PCR forced restriction enzyme digestion using the following upstream and downstream oligonucleotide primers 5^' GGC ACT GGT CTT ATA CAC CC 3^' and 5^' AAG TCA AGA GGA CTG AAC CGG 3'. The bases indicated in bold were mis-matched in the reverse primer to introduce the c-55t polymorphism sensitive MspI restriction site and in the forward primer to generate a control site to assess the completeness of restriction. The two penultimate base mismatches also increase the specificity of the subsequent PCR reaction. PCR²² buffers were 50 pM each oligonucleotide, 0.2 mmol/l of each deoxynucleotide, 7% DMSO, Bioline (London, UK) potassium chloride buffer and 0.2 units of Bioline Taq DNA polymerase in a 50 µl reaction volume. Reactions were performed using a Stratagene Robocycler (California, USA). PCR conditions were 95°C for 4 min, and 35 cycles of 95°C for 30 s, 60°C for 30 s and 72°C for 30 s. Five units of MspI (New England Biolabs Inc.) were added directly to the PCR reactions, which were incubated over-night (10-14 h) at 37°C.

High throughput electrophoresis of PCR products: Polyacrylamide gel electrophoresis was performed using microtitre-plate array diagonal gel electrophoresis (MADGE) as described previously.²³ The gels were stained with 1 µg/ml ethidium bromide and photographed under UV light. Genotypes were scored as cc, ct or tt if both alleles, one allele or neither allele cut at the -55 position with MspI, c-55 being the more common allele. Genotypes were confirmed by direct sequencing of the PCR products as described above.

Statistical Analysis: Statistical analysis was performed using SAS 6.12 for Solaris. Estimates of certain anthropometric parameters were only available in the follow-up study (AMR, PAL and VO₂). Follow-up and baseline data were available for the majority of the parameter estimates, so the dataset is repeated and unbalanced. To maximize statistical power we incorporated all of the available data into a single set using a mixed model, this treats repeated observations from the same subject, but not different subjects, as correlated.²⁴ This model was used to test the association of the UCP3 genotype with the anthropometric and biophysical outcomes, allowing for correction of confounding variables. Expression of the c-55t phenotype was assumed to be a codominant model with heterozygotes (ct) being intermediate between the two homozygous (cc, tt) phenotypes. Skewed variables were logarithmically transformed prior to parametric analysis as indicated. General linear regression analysis was performed to establish the possible effects of the c-55t polymorphism on the association between BMI and biochemical and dietary variables.

Results

Four SSCP variants were identified and sequenced from the cohort of obese children. One was intronic (intron 4; -c36t) and two exonic but silent (exon 3, t297c/Y99Y; and exon 5, c630t/Y210Y). These sequence variants have been previously reported.^6,8 A conservative mutation that encodes a conservative amino acid change in exon 3 (V102I) was detected. This was found in a 16-y-old boy of non-consanguinous Afro-Carribean descent. This subject has severe, early-onset obesity, weighing 240 kg (BMI 86 kg/m²) with no family history of obesity. This mutation has been previously identified in obese African Americans, although no convincing evidence for segregation with obesity was apparent. The gene frequency of this variant is likely to be high (~18%) in this ethnic group.³

A total of 176 out of the 838 alleles analysed from the Ely population carried the c-55t allele (21%). The distribution of alleles did not differ from the Hardy-Weinberg equilibrium (² test, P=0.9778). The genotype specific means for the anthropometic, energetic and biochemical outcomes are summarized in Table 1. Phenotypic data is presented separately for the baseline and follow-up studies.

In order to maximize statistical power, mixed models were used to analyse age- and gender-corrected data using phenotypic data combined from the baseline and follow-up study. The following model was used:

Response = intercept + A*c - 55t + B*sex + C*age

(+ D*UCP3*sex and/or E*UCP3*age if the interaction exists)

Gender: men=1, women=0.

co-dominant model: c-55t genotype,

0 = cc, 1 = ct, 2 = tt,

where A, B, C, D and E are parameter-specific coefficients.

The results of the mixed model analysis were as follows:

BMI = 23.33 - 0.734UCP3 + 0.44sex + 0.058age

BMI was significantly associated with UCP3 genotype, subjects carrying for the t-55 allele having lower BMI (P<0.0037). BMI was significantly associated with age (P=0.0012) but not gender (P>0.05) using this model.

No effect of the c-55t polymorphism on WHR, PAL, AMR, VO_2max, percentage body fat or any of the biochemical parameters measured in the whole cohort (fasting plasma glucose, triglyceride, NEFA and insulin, 30 and 120 min plasma NEFA or insulin increment was apparent (data not shown). In women, however, NEFA suppression following the glucose challenge was significantly increased in carriers of the t-55 allele:

AUC (NEFA) = 0.185 - 0.051UCP3 + 0.08age

(P = 0.026 for UCP3)

As the c-55t polymorphism is associated with BMI, we examined the effect of this polymorphism on relations between other variables that associate with BMI, specifically measures of insulin resistance (fasting, 2 h post-OGTT and AUC plasma glucose, fasting plasma insulin, NEFA and triglyceride), AMR and PAL (Table 2 and data not shown). In women the relation of fasting triglyceride and BMI was dependent on the c-55t genotype, with homozygotes for the c-55 allele showing a stronger positive correlation (Table 2). In men who were homozygous for the c-55 allele, BMI was more closely related to fasting triglyceride than heterozygotes but this effect did not reach conventional statistical significance. No effect of the c-55t polymorphism on the relations between BMI and PAL, fasting plasma insulin or NEFA was apparent.

The dietary recall information allowed us to examine the relations between percentage dietary fat intake and BMI. This relation was significantly affected by the c-55t polymorphism in men alone, c-55 homozygotes having a greater dependence of BMI on dietary fat intake.

We were also able to use the baseline and follow-up data to observe the effect of the c-55t polymorphism on longitudinal changes in the anthropometric and biochemical parameters. No significant associations were identified in this study. However, this was less powerful than the mixed model.

Discussion

Our data from sequencing the UCP3 coding sequence suggest that structural abnormalities of UCP3 are unlikely to be a major cause of severe juvenile obesity. This data is in agreement with other studies in both Caucasian juveniles^6,8 and adult African Americans.⁷

We present data showing that the UCP3 c-55t promoter polymorphism associates with reduced BMI in a free-living Caucasian population. The nature of the association of this polymorphism with body mass and composition is likely to be complex. This polymorphism is not silent in Pima Indians, as subjects who carry the t-55 allele have been shown to have significantly higher UCP3 mRNA concentration than c-55 homozygotes.¹⁰ There is also data from Pimas showing that a decrease in full length UCP3 message associates with a reduction in BMI, which is likely to be due to an observed reduction in fat oxidation rates.³ The simplest explanation for our findings is that the c-55t allele is associated with a reduction in UCP3 message, which, if similar to the situation in Pimas, is associated with increased fat oxidation and reduced BMI. Unfortunately, as yet, we have no data on rates of fat oxidation in this cohort. The situation, however, is complicated by the existence of two forms of the UCP3 message, a short and a long form.²⁵ Whilst the short form is equally,²⁶ or even more,²⁷ able to uncouple yeast mitochondria, it is incorporated into the mitochondria of cultured cells more slowly than the longer form.²⁸ Proof of this hypothesis then requires knowledge of direct effects of the c-55t polymorphism on long and short UCP3 message transcription efficiency and relations between UCP3 activity and message concentration in vivo.²⁹

Otabe et al¹¹ published data on the c-55t polymorphism in cohorts of obese and non-obese French subjects that show contradictory associations. Whether either or both of these studies represent type I statistical errors awaits the result of further epidemiological and biochemical studies on the effects of this polymorphism.

The effect on BMI in this cohort is not reflected in changes in WHR or percentage body fat. In both men and women WHR was less in c-55t homozygotes than in carriers of the c-55 allele both at baseline and follow-up. Percentage body fat also showed a similar trend at follow-up (data not available at baseline) but none of these changes reached significance. This lack of significance may be due to insufficient power to detect changes in these parameters. None of the other anthropometric indices measured (PAL, VO_2max or AMR) were associated with UCP3 genotype.

There is considerable data to suggest that UCP3 message levels correlate with NEFA concentrations.^30,31,32,33 This study gave us the ideal opportunity to study possible associations of this polymorphism with fasting plasma NEFA under conditions where skeletal muscle will be predominantly metabolizing NEFA. No such associations were apparent. If UCP3 is a fatty acid transporter then the putative up-regulation of its mRNA has no effect on fasting NEFA in the Ely cohort.

Several lines of evidence suggest that NEFA levels may not be the prime regulator of UCP3 action and that reduced muscle glucose uptake may be the dominant effect.^1,34 No associations of the UCP3 c-55t genotype with fasting plasma insulin or glucose were evident from these data. We also saw no associations with 2 h post-OGTT plasma glucose or AUC glucose, suggesting no major associations of this locus with insulin resistance. NEFA suppression following the glucose challenge was, however, significantly increased in female carriers of the t-55 allele. This affect was not apparent in men who, in this cohort, are more insulin resistant.³⁵ Female carriers of the t-55 allele then seem better able to switch from fatty acid to glucose metabolism following an insulin surge. The increased NEFA suppression associated with the t-55 genotype may be exposed by the greater insulin sensitivity in women.

The contribution of the c-55t polymorphism to the distribution of BMI within the Ely cohort is small. Other environmental or genetic factors are likely to affect the expressivity of this polymorphism. Probable candidates would affect carbohydrate or fat metabolism. We were able to examine the interaction of this polymorphism with intermediate phenotypes such as insulin resistance and fasting NEFA or the environmental factor, dietary fat intake. Carriers of the c-55t allele had a weaker positive association of BMI with fasting plasma triglyceride. For a given fasting plasma triglyceride c-55t heterozygotes are likely to have a lower BMI. These differences in BMI are most apparent at high fasting plasma triglyceride concentrations. High fasting plasma triglyceride concentration is associated with higher fat oxidation rates and this may help to expose the c-55t phenotype. Again, proof of this hypothesis requires measurement of fat oxidation rates in these individuals. Data taken from the men also suggest that BMI in carriers of the c-55t allele is less dependent on dietary fat. As this is not found in women the significance of this result is questionable and may reflect inherent problems in collection of this type of data.

In conclusion we have shown that variation at the UCP3 gene is unlikely to be a common monogenic cause of severe obesity. However in a Caucasian population, BMI was significantly associated, in a co-dominant manner, with the common UCP3 c-55t polymorphism, carriers of the less common allele having a lower BMI. Our data also suggests that the effects of this polymorphism may be particularly pronounced in subjects with higher fasting plasma lipids.

1 Samec S, Seydoux J, Dulloo AG. Post-starvation gene expression of skeletal muscle uncoupling protein 2 and uncoupling protein 3 in response to dietary fat levels and fatty acid composition: a link with insulin resistance. Diabetes 1999; 48: 436-441, MEDLINE

2 Ricquier D, Bouillaud F. The uncoupling protein homologues: UCP1, UCP2, UCP3, StUCP and AtUCP. Biochem J 2000; 345: 161-179, Article MEDLINE

3 Argyropoulos G, Brown AM, Willi SM, Zhu J, He Y, Reitman M, Gevao SM, Spruill I, Garvey WT. Effects of mutations in the human uncoupling protein 3 gene on the respiratory quotient and fat oxidation in severe obesity and type 2 diabetes. J Clin Invest 1998; 102: 1345-1351, MEDLINE

4 Brown AM, Willi SM, Argyropoulos G, Garvey WT. A novel missense mutation, R70W, in the human uncoupling protein 3 gene in a family with type 2 diabetes. Hum Mut 1999; 13: 506,

5 Brown AM, Dolan JW, Willi SM, Garvey WT, Argyropoulos G. Endogenous mutations in human uncoupling protein 3 alter its functional properties. FEBS Lett 1999; 464: 189-193, MEDLINE

6 Urhammer SA, Dalgaard LT, Sorensen TI, Tybjaerg-Hansen A, Echwald SM, Andersen T, Clausen JO, Pedersen O. Organisation of the coding exons and mutational screening of the uncoupling protein 3 gene in subjects with juvenile-onset obesity. Diabetologia 1998; 41: 241-244, MEDLINE

7 Chung WK, Luke A, Cooper RS, Rotini C, Vidal-Puig A, Rosenbaum M, Chua M, Solanes G, Zheng M, Zhao L, LeDuc C, Eisberg A, Chu F, Murphy E, Schreier M, Aronne L, Caprio S, Kahle B, Gordon D, Leal SM, Goldsmith R, Andreu AL, Bruno C, DiMauro S, Leibel RL. Genetic and physiologic analysis of the role of uncoupling protein 3 in human energy homeostasis. Diabetes 1999; 48: 1890-1895, MEDLINE

8 Otabe S, Clement K, Dubois S, Lepretre F, Pelloux V, Leibel R, Chung W, Boutin, P, Guy-Grand B, Froguel P, Vasseur, F. Mutation screening and association studies of the human uncoupling protein 3 gene in normoglycaemic and diabetic morbidly obese patients. Diabetes 1999; 48: 206-208, MEDLINE

9 Saker PJ, Cassel PG, Huxtable SJ, Kousta E, Hattersley AT, Frayling TM, Walker M, Kopelman PG, Ramachandran A, Snehelatha C, McCarthy MI, Hitman GA. Diabetologia (in press).

10 Schrauwen P, Xia J, Walder K, Snitker S, Ravussin E. A novel polymorphism in the proximal UCP3 promoter region: effect on skeletal muscle UCP3 mRNA expression and obesity in male non-diabetic Pima Indians. Int J Obes Relat Metab Disord 1999; 23: 1242-1245, MEDLINE

11 Otabe S, Clement K, Dina C, Pelloux V, Guy-Grand B, Froguel P, Vasseur F. A genetic variation in the 5^' flanking region of the UCP3 gene is associated with body mass index in humans in interaction with physical activity. Diabetologia 2000; 43: 245-249, MEDLINE

12 Acin A, Rodriguez M, Rique H, Canet E, Boutin JA, Galizzi JP. Cloning and characterization of the 5^' flanking region of the human uncoupling protein 3 (UCP3) gene. Biochem Biophys Res Commun 1999; 258: 278-283, MEDLINE

13 Williams DR, Wareham NJ, Brown DC, Byrne CD, Clark PM, Cox BD, Cox LJ, Day NE, Hales CN, Palmer CR. Undiagnosed glucose intolerance in the community: the Isle of Ely Diabetes Project. Diabetic Med 1995; 12: 30-35, MEDLINE

14 Wareham NJ, Byrne CD, Williams DR, Day NE, Hales CN. Fasting proinsulin concentrations predict the development of type 2 diabetes. Diabetes Care 1999; 22: 262-270, MEDLINE

15 Wareham NJ, Hennings SJ, Byrne CD, Hales CN, Prentice AM, Day NE. A quantitative analysis of the relationship between habitual energy expenditure, fitness and the metabolic cardiovascular syndrome. Br J Nutr 1998; 80: 235-241, MEDLINE

16 Wareham NJ, Wong MY, Day NE. Glucose intolerance and physical inactivity: the relative importance of low habitual energy expenditure and cardiorespiratory fitness. Am J Epidemiol 2000; 152: 32-139, MEDLINE

17 Halsall DJ, Martensz ND, Luan J, Wareham NJ, Hales CN, Byrne CD. A common apolipoprotein B signal peptide polymorphism modifies the relation between plasma non-esterified fatty acid and triglyceride concentration in men. Atherosclerosis 2000; 152: 9-17, MEDLINE

18 Phillips DI, Clark PM, Hales CN, Osmond C. Understanding oral glucose tolerance: comparison of glucose or insulin measurements during the oral glucose tolerance test with specific measurements of insulin resistance and insulin secretion. Diabetic Med 1994; 11: 286, MEDLINE

19 Byrne CD, Wareham NJ, Day NE, McLeish R, Williams DR, Hales CN. Decreased non-esterified fatty acid suppression and features of the insulin resistance syndrome occur in a sub-group of individuals with normal glucose tolerance. Diabetologia 1995; 38: 1358, MEDLINE

20 Miller SA, Dykes DD, Polesky HF. A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 1988; 16: 1215, MEDLINE

21 Krook A, Kumar S, Laing I, Boulton AJ, Wass JA, O'Rahilly S. Molecular scanning of the insulin receptor gene in syndromes of insulin resistance. Diabetes 1994; 43: 357-368, MEDLINE

22 Mullis K, Faloona F, Scharf S, Saiki R, Horn G, Erlich H. Specific enzymatic amplification of DNA in vitro: the polymerase chain reaction. Cold Spring Harb Symp Quant Biol 1986; 51: 263, MEDLINE

23 Day IN, Humphries SE, Richards S, Norton D, Reid M. High-throughput genotyping using horizontal polyacrylamide gels with wells arranged for microplate array diagonal gel electrophoresis MADGE. Biotechniques 1995; 19: 830, MEDLINE

24 Littell RC, Milliken GA, Stroup WW, Wolfinger RD. SAS (R) System for Mixed Models. SAS: Cary, NC, 1999,

25 Solanes G, Vidal-Puig A, Grujic D, Flier JS, Lowell BB. The human uncoupling protein-3 gene. Genomic structure, chromosomal localization, and genetic basis for short and long form transcripts. J Biol Chem 1997; 272: 25433-25436, MEDLINE

26 Brown AM, Dolan JW, Willi SM, Garvey WT, Argyropoulos G. Endogenous mutations in human uncoupling protein 3 alter its functional properties. FEBS Lett 1999; 464: 189-193, MEDLINE

27 Hinz W, Gruninger S, De Pover A, Chiesi M. Properties of the human long and short isoforms of the uncoupling protein-3 expressed in yeast cells. FEBS Lett 1999; 462: 411-415, MEDLINE

28 Renold A, Koehler CM, Murphy MP. Mitochondrial import of the long and short isoforms of human uncoupling protein 3. FEBS Lett 2000; 465: 135-140, MEDLINE

29 Silva JE. The physiological role of the novel uncoupling proteins: view from the chair. Int J Obes Relat Metab Disord 1999; 23(Suppl 6): S72-S74, MEDLINE

30 Giacobino JP. Effects of dietary deprivation, obesity and exercise on UCP3 mRNA levels. Int J Obes Relat Metab Disord 1999; 23(Suppl 6): S60-S63, MEDLINE

31 Samec S, Seydoux J, Dulloo AG. Skeletal muscle UCP3 and UCP2 gene expression in response to inhibition of free fatty acid flux through mitochondrial beta-oxidation. Pflügers Arch 1999; 438: 452-457, Article MEDLINE

32 Weigle DS, Selfridge LE, Schwartz MW, Seeley RJ, Cummings DE, Havel PJ, Kuijper JL, BeltrandelRio H. Elevated free fatty acids induce uncoupling protein 3 expression in muscle: a potential explanation for the effect of fasting. Diabetes 1998; 47: 298-302, MEDLINE

33 Hwang CS, Lane MD. Up-regulation of uncoupling protein-3 by fatty acid in C2C12 myotubes. Biochem Biophys Res Commun 1999; 258: 464-469, Article MEDLINE

34 Tsuboyama-Kasaoka N, Tsunoda N, Maruyama K, Takahashi M, Kim H, Cooke DW, Lane MD, Ezaki O. Overexpression of GLUT4 in mice causes up-regulation of UCP3 mRNA in skeletal muscle. Biochem Biophys Res Commun 1999; 258: 187-193, MEDLINE

35 Byrne CD, Wareham NJ, Day NE, McLeish R, Williams DR, Hales CN. Decreased non-esterified fatty acid suppression and features of the insulin resistance syndrome occur in a sub-group of individuals with normal glucose tolerance. Diabetologia 1995; 38: 1358-1366, MEDLINE

Table 1 Genotype specific means for anthropological and biochemical data

Table 2 Effects of c-55t polymorphism on the regression coefficients between body mass index and associated parameters